mirror of
https://github.com/netbirdio/netbird.git
synced 2026-04-02 07:33:52 -04:00
Compare commits
8 Commits
feature/di
...
sync-clien
| Author | SHA1 | Date | |
|---|---|---|---|
|
|
72513d7522 | ||
|
|
a1f1bf1f19 | ||
|
|
b5dec3df39 | ||
|
|
20f5f00635 | ||
|
|
fc141cf3a3 | ||
|
|
d0c65fa08e | ||
|
|
f241bfa339 | ||
|
|
4b2cd97d5f |
@@ -1,15 +1,15 @@
|
||||
FROM golang:1.25-bookworm
|
||||
FROM golang:1.23-bullseye
|
||||
|
||||
RUN apt-get update && export DEBIAN_FRONTEND=noninteractive \
|
||||
&& apt-get -y install --no-install-recommends\
|
||||
gettext-base=0.21-12 \
|
||||
iptables=1.8.9-2 \
|
||||
libgl1-mesa-dev=22.3.6-1+deb12u1 \
|
||||
xorg-dev=1:7.7+23 \
|
||||
libayatana-appindicator3-dev=0.5.92-1 \
|
||||
gettext-base=0.21-4 \
|
||||
iptables=1.8.7-1 \
|
||||
libgl1-mesa-dev=20.3.5-1 \
|
||||
xorg-dev=1:7.7+22 \
|
||||
libayatana-appindicator3-dev=0.5.5-2+deb11u2 \
|
||||
&& apt-get clean \
|
||||
&& rm -rf /var/lib/apt/lists/* \
|
||||
&& go install -v golang.org/x/tools/gopls@latest
|
||||
&& go install -v golang.org/x/tools/gopls@v0.18.1
|
||||
|
||||
|
||||
WORKDIR /app
|
||||
|
||||
4
.github/workflows/golang-test-freebsd.yml
vendored
4
.github/workflows/golang-test-freebsd.yml
vendored
@@ -25,7 +25,7 @@ jobs:
|
||||
release: "14.2"
|
||||
prepare: |
|
||||
pkg install -y curl pkgconf xorg
|
||||
GO_TARBALL="go1.25.3.freebsd-amd64.tar.gz"
|
||||
GO_TARBALL="go1.24.10.freebsd-amd64.tar.gz"
|
||||
GO_URL="https://go.dev/dl/$GO_TARBALL"
|
||||
curl -vLO "$GO_URL"
|
||||
tar -C /usr/local -vxzf "$GO_TARBALL"
|
||||
@@ -39,7 +39,7 @@ jobs:
|
||||
# check all component except management, since we do not support management server on freebsd
|
||||
time go test -timeout 1m -failfast ./base62/...
|
||||
# NOTE: without -p1 `client/internal/dns` will fail because of `listen udp4 :33100: bind: address already in use`
|
||||
time go test -timeout 8m -failfast -v -p 1 ./client/...
|
||||
time go test -timeout 8m -failfast -p 1 ./client/...
|
||||
time go test -timeout 1m -failfast ./dns/...
|
||||
time go test -timeout 1m -failfast ./encryption/...
|
||||
time go test -timeout 1m -failfast ./formatter/...
|
||||
|
||||
4
.github/workflows/golang-test-linux.yml
vendored
4
.github/workflows/golang-test-linux.yml
vendored
@@ -200,7 +200,7 @@ jobs:
|
||||
-e GOCACHE=${CONTAINER_GOCACHE} \
|
||||
-e GOMODCACHE=${CONTAINER_GOMODCACHE} \
|
||||
-e CONTAINER=${CONTAINER} \
|
||||
golang:1.25-alpine \
|
||||
golang:1.24-alpine \
|
||||
sh -c ' \
|
||||
apk update; apk add --no-cache \
|
||||
ca-certificates iptables ip6tables dbus dbus-dev libpcap-dev build-base; \
|
||||
@@ -259,7 +259,7 @@ jobs:
|
||||
CGO_ENABLED=1 GOARCH=${{ matrix.arch }} \
|
||||
go test ${{ matrix.raceFlag }} \
|
||||
-exec 'sudo' \
|
||||
-timeout 10m -p 1 ./relay/... ./shared/relay/...
|
||||
-timeout 10m ./relay/... ./shared/relay/...
|
||||
|
||||
test_signal:
|
||||
name: "Signal / Unit"
|
||||
|
||||
7
.github/workflows/golangci-lint.yml
vendored
7
.github/workflows/golangci-lint.yml
vendored
@@ -52,10 +52,7 @@ jobs:
|
||||
if: matrix.os == 'ubuntu-latest'
|
||||
run: sudo apt update && sudo apt install -y -q libgtk-3-dev libayatana-appindicator3-dev libgl1-mesa-dev xorg-dev libpcap-dev
|
||||
- name: golangci-lint
|
||||
uses: golangci/golangci-lint-action@4afd733a84b1f43292c63897423277bb7f4313a9 # v8.0.0
|
||||
uses: golangci/golangci-lint-action@v4
|
||||
with:
|
||||
version: latest
|
||||
skip-cache: true
|
||||
skip-save-cache: true
|
||||
cache-invalidation-interval: 0
|
||||
args: --timeout=12m
|
||||
args: --timeout=12m --out-format colored-line-number
|
||||
|
||||
96
.github/workflows/release.yml
vendored
96
.github/workflows/release.yml
vendored
@@ -9,7 +9,7 @@ on:
|
||||
pull_request:
|
||||
|
||||
env:
|
||||
SIGN_PIPE_VER: "v0.1.0"
|
||||
SIGN_PIPE_VER: "v0.0.23"
|
||||
GORELEASER_VER: "v2.3.2"
|
||||
PRODUCT_NAME: "NetBird"
|
||||
COPYRIGHT: "NetBird GmbH"
|
||||
@@ -19,100 +19,6 @@ concurrency:
|
||||
cancel-in-progress: true
|
||||
|
||||
jobs:
|
||||
release_freebsd_port:
|
||||
name: "FreeBSD Port / Build & Test"
|
||||
runs-on: ubuntu-22.04
|
||||
steps:
|
||||
- name: Checkout
|
||||
uses: actions/checkout@v4
|
||||
|
||||
- name: Generate FreeBSD port diff
|
||||
run: bash release_files/freebsd-port-diff.sh
|
||||
|
||||
- name: Generate FreeBSD port issue body
|
||||
run: bash release_files/freebsd-port-issue-body.sh
|
||||
|
||||
- name: Check if diff was generated
|
||||
id: check_diff
|
||||
run: |
|
||||
if ls netbird-*.diff 1> /dev/null 2>&1; then
|
||||
echo "diff_exists=true" >> $GITHUB_OUTPUT
|
||||
else
|
||||
echo "diff_exists=false" >> $GITHUB_OUTPUT
|
||||
echo "No diff file generated (port may already be up to date)"
|
||||
fi
|
||||
|
||||
- name: Extract version
|
||||
if: steps.check_diff.outputs.diff_exists == 'true'
|
||||
id: version
|
||||
run: |
|
||||
VERSION=$(ls netbird-*.diff | sed 's/netbird-\(.*\)\.diff/\1/')
|
||||
echo "version=$VERSION" >> $GITHUB_OUTPUT
|
||||
echo "Generated files for version: $VERSION"
|
||||
cat netbird-*.diff
|
||||
|
||||
- name: Test FreeBSD port
|
||||
if: steps.check_diff.outputs.diff_exists == 'true'
|
||||
uses: vmactions/freebsd-vm@v1
|
||||
with:
|
||||
usesh: true
|
||||
copyback: false
|
||||
release: "15.0"
|
||||
prepare: |
|
||||
# Install required packages
|
||||
pkg install -y git curl portlint go
|
||||
|
||||
# Install Go for building
|
||||
GO_TARBALL="go1.25.5.freebsd-amd64.tar.gz"
|
||||
GO_URL="https://go.dev/dl/$GO_TARBALL"
|
||||
curl -LO "$GO_URL"
|
||||
tar -C /usr/local -xzf "$GO_TARBALL"
|
||||
|
||||
# Clone ports tree (shallow, only what we need)
|
||||
git clone --depth 1 --filter=blob:none https://git.FreeBSD.org/ports.git /usr/ports
|
||||
cd /usr/ports
|
||||
|
||||
run: |
|
||||
set -e -x
|
||||
export PATH=$PATH:/usr/local/go/bin
|
||||
|
||||
# Find the diff file
|
||||
echo "Finding diff file..."
|
||||
DIFF_FILE=$(find $PWD -name "netbird-*.diff" -type f 2>/dev/null | head -1)
|
||||
echo "Found: $DIFF_FILE"
|
||||
|
||||
if [[ -z "$DIFF_FILE" ]]; then
|
||||
echo "ERROR: Could not find diff file"
|
||||
find ~ -name "*.diff" -type f 2>/dev/null || true
|
||||
exit 1
|
||||
fi
|
||||
|
||||
# Apply the generated diff from /usr/ports (diff has a/security/netbird/... paths)
|
||||
cd /usr/ports
|
||||
patch -p1 -V none < "$DIFF_FILE"
|
||||
|
||||
# Show patched Makefile
|
||||
version=$(cat security/netbird/Makefile | grep -E '^DISTVERSION=' | awk '{print $NF}')
|
||||
|
||||
cd /usr/ports/security/netbird
|
||||
export BATCH=yes
|
||||
make package
|
||||
pkg add ./work/pkg/netbird-*.pkg
|
||||
|
||||
netbird version | grep "$version"
|
||||
|
||||
echo "FreeBSD port test completed successfully!"
|
||||
|
||||
- name: Upload FreeBSD port files
|
||||
if: steps.check_diff.outputs.diff_exists == 'true'
|
||||
uses: actions/upload-artifact@v4
|
||||
with:
|
||||
name: freebsd-port-files
|
||||
path: |
|
||||
./netbird-*-issue.txt
|
||||
./netbird-*.diff
|
||||
retention-days: 30
|
||||
|
||||
release:
|
||||
runs-on: ubuntu-latest-m
|
||||
env:
|
||||
|
||||
@@ -243,7 +243,6 @@ jobs:
|
||||
working-directory: infrastructure_files/artifacts
|
||||
run: |
|
||||
sleep 30
|
||||
docker compose logs
|
||||
docker compose exec management ls -l /var/lib/netbird/ | grep -i GeoLite2-City_[0-9]*.mmdb
|
||||
docker compose exec management ls -l /var/lib/netbird/ | grep -i geonames_[0-9]*.db
|
||||
|
||||
|
||||
13
.github/workflows/wasm-build-validation.yml
vendored
13
.github/workflows/wasm-build-validation.yml
vendored
@@ -14,9 +14,6 @@ jobs:
|
||||
js_lint:
|
||||
name: "JS / Lint"
|
||||
runs-on: ubuntu-latest
|
||||
env:
|
||||
GOOS: js
|
||||
GOARCH: wasm
|
||||
steps:
|
||||
- name: Checkout repository
|
||||
uses: actions/checkout@v4
|
||||
@@ -27,14 +24,16 @@ jobs:
|
||||
- name: Install dependencies
|
||||
run: sudo apt update && sudo apt install -y -q libgtk-3-dev libayatana-appindicator3-dev libgl1-mesa-dev xorg-dev libpcap-dev
|
||||
- name: Install golangci-lint
|
||||
uses: golangci/golangci-lint-action@4afd733a84b1f43292c63897423277bb7f4313a9 # v8.0.0
|
||||
uses: golangci/golangci-lint-action@d6238b002a20823d52840fda27e2d4891c5952dc
|
||||
with:
|
||||
version: latest
|
||||
install-mode: binary
|
||||
skip-cache: true
|
||||
skip-save-cache: true
|
||||
cache-invalidation-interval: 0
|
||||
working-directory: ./client
|
||||
skip-pkg-cache: true
|
||||
skip-build-cache: true
|
||||
- name: Run golangci-lint for WASM
|
||||
run: |
|
||||
GOOS=js GOARCH=wasm golangci-lint run --timeout=12m --out-format colored-line-number ./client/...
|
||||
continue-on-error: true
|
||||
|
||||
js_build:
|
||||
|
||||
1
.gitignore
vendored
1
.gitignore
vendored
@@ -31,4 +31,3 @@ infrastructure_files/setup-*.env
|
||||
.DS_Store
|
||||
vendor/
|
||||
/netbird
|
||||
client/netbird-electron/
|
||||
|
||||
257
.golangci.yaml
257
.golangci.yaml
@@ -1,124 +1,139 @@
|
||||
version: "2"
|
||||
linters:
|
||||
default: none
|
||||
enable:
|
||||
- bodyclose
|
||||
- dupword
|
||||
- durationcheck
|
||||
- errcheck
|
||||
- forbidigo
|
||||
- gocritic
|
||||
- gosec
|
||||
- govet
|
||||
- ineffassign
|
||||
- mirror
|
||||
- misspell
|
||||
- nilerr
|
||||
- nilnil
|
||||
- predeclared
|
||||
- revive
|
||||
- sqlclosecheck
|
||||
- staticcheck
|
||||
- unused
|
||||
- wastedassign
|
||||
settings:
|
||||
errcheck:
|
||||
check-type-assertions: false
|
||||
gocritic:
|
||||
disabled-checks:
|
||||
- commentFormatting
|
||||
- captLocal
|
||||
- deprecatedComment
|
||||
gosec:
|
||||
includes:
|
||||
- G101
|
||||
- G103
|
||||
- G104
|
||||
- G106
|
||||
- G108
|
||||
- G109
|
||||
- G110
|
||||
- G111
|
||||
- G201
|
||||
- G202
|
||||
- G203
|
||||
- G301
|
||||
- G302
|
||||
- G303
|
||||
- G304
|
||||
- G305
|
||||
- G306
|
||||
- G307
|
||||
- G403
|
||||
- G502
|
||||
- G503
|
||||
- G504
|
||||
- G601
|
||||
- G602
|
||||
govet:
|
||||
enable:
|
||||
- nilness
|
||||
enable-all: false
|
||||
revive:
|
||||
rules:
|
||||
- name: exported
|
||||
arguments:
|
||||
- checkPrivateReceivers
|
||||
- sayRepetitiveInsteadOfStutters
|
||||
severity: warning
|
||||
disabled: false
|
||||
exclusions:
|
||||
generated: lax
|
||||
presets:
|
||||
- comments
|
||||
- common-false-positives
|
||||
- legacy
|
||||
- std-error-handling
|
||||
run:
|
||||
# Timeout for analysis, e.g. 30s, 5m.
|
||||
# Default: 1m
|
||||
timeout: 6m
|
||||
|
||||
# This file contains only configs which differ from defaults.
|
||||
# All possible options can be found here https://github.com/golangci/golangci-lint/blob/master/.golangci.reference.yml
|
||||
linters-settings:
|
||||
errcheck:
|
||||
# Report about not checking of errors in type assertions: `a := b.(MyStruct)`.
|
||||
# Such cases aren't reported by default.
|
||||
# Default: false
|
||||
check-type-assertions: false
|
||||
|
||||
gosec:
|
||||
includes:
|
||||
- G101 # Look for hard coded credentials
|
||||
#- G102 # Bind to all interfaces
|
||||
- G103 # Audit the use of unsafe block
|
||||
- G104 # Audit errors not checked
|
||||
- G106 # Audit the use of ssh.InsecureIgnoreHostKey
|
||||
#- G107 # Url provided to HTTP request as taint input
|
||||
- G108 # Profiling endpoint automatically exposed on /debug/pprof
|
||||
- G109 # Potential Integer overflow made by strconv.Atoi result conversion to int16/32
|
||||
- G110 # Potential DoS vulnerability via decompression bomb
|
||||
- G111 # Potential directory traversal
|
||||
#- G112 # Potential slowloris attack
|
||||
- G113 # Usage of Rat.SetString in math/big with an overflow (CVE-2022-23772)
|
||||
#- G114 # Use of net/http serve function that has no support for setting timeouts
|
||||
- G201 # SQL query construction using format string
|
||||
- G202 # SQL query construction using string concatenation
|
||||
- G203 # Use of unescaped data in HTML templates
|
||||
#- G204 # Audit use of command execution
|
||||
- G301 # Poor file permissions used when creating a directory
|
||||
- G302 # Poor file permissions used with chmod
|
||||
- G303 # Creating tempfile using a predictable path
|
||||
- G304 # File path provided as taint input
|
||||
- G305 # File traversal when extracting zip/tar archive
|
||||
- G306 # Poor file permissions used when writing to a new file
|
||||
- G307 # Poor file permissions used when creating a file with os.Create
|
||||
#- G401 # Detect the usage of DES, RC4, MD5 or SHA1
|
||||
#- G402 # Look for bad TLS connection settings
|
||||
- G403 # Ensure minimum RSA key length of 2048 bits
|
||||
#- G404 # Insecure random number source (rand)
|
||||
#- G501 # Import blocklist: crypto/md5
|
||||
- G502 # Import blocklist: crypto/des
|
||||
- G503 # Import blocklist: crypto/rc4
|
||||
- G504 # Import blocklist: net/http/cgi
|
||||
#- G505 # Import blocklist: crypto/sha1
|
||||
- G601 # Implicit memory aliasing of items from a range statement
|
||||
- G602 # Slice access out of bounds
|
||||
|
||||
gocritic:
|
||||
disabled-checks:
|
||||
- commentFormatting
|
||||
- captLocal
|
||||
- deprecatedComment
|
||||
|
||||
govet:
|
||||
# Enable all analyzers.
|
||||
# Default: false
|
||||
enable-all: false
|
||||
enable:
|
||||
- nilness
|
||||
|
||||
revive:
|
||||
rules:
|
||||
- linters:
|
||||
- forbidigo
|
||||
path: management/cmd/root\.go
|
||||
- linters:
|
||||
- forbidigo
|
||||
path: signal/cmd/root\.go
|
||||
- linters:
|
||||
- unused
|
||||
path: sharedsock/filter\.go
|
||||
- linters:
|
||||
- unused
|
||||
path: client/firewall/iptables/rule\.go
|
||||
- linters:
|
||||
- gosec
|
||||
- mirror
|
||||
path: test\.go
|
||||
- linters:
|
||||
- nilnil
|
||||
path: mock\.go
|
||||
- linters:
|
||||
- staticcheck
|
||||
text: grpc.DialContext is deprecated
|
||||
- linters:
|
||||
- staticcheck
|
||||
text: grpc.WithBlock is deprecated
|
||||
- linters:
|
||||
- staticcheck
|
||||
text: "QF1001"
|
||||
- linters:
|
||||
- staticcheck
|
||||
text: "QF1008"
|
||||
- linters:
|
||||
- staticcheck
|
||||
text: "QF1012"
|
||||
paths:
|
||||
- third_party$
|
||||
- builtin$
|
||||
- examples$
|
||||
- name: exported
|
||||
severity: warning
|
||||
disabled: false
|
||||
arguments:
|
||||
- "checkPrivateReceivers"
|
||||
- "sayRepetitiveInsteadOfStutters"
|
||||
tenv:
|
||||
# The option `all` will run against whole test files (`_test.go`) regardless of method/function signatures.
|
||||
# Otherwise, only methods that take `*testing.T`, `*testing.B`, and `testing.TB` as arguments are checked.
|
||||
# Default: false
|
||||
all: true
|
||||
|
||||
linters:
|
||||
disable-all: true
|
||||
enable:
|
||||
## enabled by default
|
||||
- errcheck # checking for unchecked errors, these unchecked errors can be critical bugs in some cases
|
||||
- gosimple # specializes in simplifying a code
|
||||
- govet # reports suspicious constructs, such as Printf calls whose arguments do not align with the format string
|
||||
- ineffassign # detects when assignments to existing variables are not used
|
||||
- staticcheck # is a go vet on steroids, applying a ton of static analysis checks
|
||||
- tenv # Tenv is analyzer that detects using os.Setenv instead of t.Setenv since Go1.17.
|
||||
- typecheck # like the front-end of a Go compiler, parses and type-checks Go code
|
||||
- unused # checks for unused constants, variables, functions and types
|
||||
## disable by default but the have interesting results so lets add them
|
||||
- bodyclose # checks whether HTTP response body is closed successfully
|
||||
- dupword # dupword checks for duplicate words in the source code
|
||||
- durationcheck # durationcheck checks for two durations multiplied together
|
||||
- forbidigo # forbidigo forbids identifiers
|
||||
- gocritic # provides diagnostics that check for bugs, performance and style issues
|
||||
- gosec # inspects source code for security problems
|
||||
- mirror # mirror reports wrong mirror patterns of bytes/strings usage
|
||||
- misspell # misspess finds commonly misspelled English words in comments
|
||||
- nilerr # finds the code that returns nil even if it checks that the error is not nil
|
||||
- nilnil # checks that there is no simultaneous return of nil error and an invalid value
|
||||
- predeclared # predeclared finds code that shadows one of Go's predeclared identifiers
|
||||
- revive # Fast, configurable, extensible, flexible, and beautiful linter for Go. Drop-in replacement of golint.
|
||||
- sqlclosecheck # checks that sql.Rows and sql.Stmt are closed
|
||||
# - thelper # thelper detects Go test helpers without t.Helper() call and checks the consistency of test helpers.
|
||||
- wastedassign # wastedassign finds wasted assignment statements
|
||||
issues:
|
||||
# Maximum count of issues with the same text.
|
||||
# Set to 0 to disable.
|
||||
# Default: 3
|
||||
max-same-issues: 5
|
||||
formatters:
|
||||
exclusions:
|
||||
generated: lax
|
||||
paths:
|
||||
- third_party$
|
||||
- builtin$
|
||||
- examples$
|
||||
|
||||
exclude-rules:
|
||||
# allow fmt
|
||||
- path: management/cmd/root\.go
|
||||
linters: forbidigo
|
||||
- path: signal/cmd/root\.go
|
||||
linters: forbidigo
|
||||
- path: sharedsock/filter\.go
|
||||
linters:
|
||||
- unused
|
||||
- path: client/firewall/iptables/rule\.go
|
||||
linters:
|
||||
- unused
|
||||
- path: test\.go
|
||||
linters:
|
||||
- mirror
|
||||
- gosec
|
||||
- path: mock\.go
|
||||
linters:
|
||||
- nilnil
|
||||
# Exclude specific deprecation warnings for grpc methods
|
||||
- linters:
|
||||
- staticcheck
|
||||
text: "grpc.DialContext is deprecated"
|
||||
- linters:
|
||||
- staticcheck
|
||||
text: "grpc.WithBlock is deprecated"
|
||||
|
||||
@@ -713,10 +713,8 @@ checksum:
|
||||
extra_files:
|
||||
- glob: ./infrastructure_files/getting-started-with-zitadel.sh
|
||||
- glob: ./release_files/install.sh
|
||||
- glob: ./infrastructure_files/getting-started.sh
|
||||
|
||||
release:
|
||||
extra_files:
|
||||
- glob: ./infrastructure_files/getting-started-with-zitadel.sh
|
||||
- glob: ./release_files/install.sh
|
||||
- glob: ./infrastructure_files/getting-started.sh
|
||||
|
||||
11
README.md
11
README.md
@@ -38,11 +38,6 @@
|
||||
|
||||
</strong>
|
||||
<br>
|
||||
<strong>
|
||||
🚀 <a href="https://careers.netbird.io">We are hiring! Join us at careers.netbird.io</a>
|
||||
</strong>
|
||||
<br>
|
||||
<br>
|
||||
<a href="https://registry.terraform.io/providers/netbirdio/netbird/latest">
|
||||
New: NetBird terraform provider
|
||||
</a>
|
||||
@@ -90,7 +85,7 @@ Follow the [Advanced guide with a custom identity provider](https://docs.netbird
|
||||
|
||||
**Infrastructure requirements:**
|
||||
- A Linux VM with at least **1CPU** and **2GB** of memory.
|
||||
- The VM should be publicly accessible on TCP ports **80** and **443** and UDP port: **3478**.
|
||||
- The VM should be publicly accessible on TCP ports **80** and **443** and UDP ports: **3478**, **49152-65535**.
|
||||
- **Public domain** name pointing to the VM.
|
||||
|
||||
**Software requirements:**
|
||||
@@ -103,7 +98,7 @@ Follow the [Advanced guide with a custom identity provider](https://docs.netbird
|
||||
**Steps**
|
||||
- Download and run the installation script:
|
||||
```bash
|
||||
export NETBIRD_DOMAIN=netbird.example.com; curl -fsSL https://github.com/netbirdio/netbird/releases/latest/download/getting-started.sh | bash
|
||||
export NETBIRD_DOMAIN=netbird.example.com; curl -fsSL https://github.com/netbirdio/netbird/releases/latest/download/getting-started-with-zitadel.sh | bash
|
||||
```
|
||||
- Once finished, you can manage the resources via `docker-compose`
|
||||
|
||||
@@ -118,7 +113,7 @@ export NETBIRD_DOMAIN=netbird.example.com; curl -fsSL https://github.com/netbird
|
||||
[Coturn](https://github.com/coturn/coturn) is the one that has been successfully used for STUN and TURN in NetBird setups.
|
||||
|
||||
<p float="left" align="middle">
|
||||
<img src="https://docs.netbird.io/docs-static/img/about-netbird/high-level-dia.png" width="700"/>
|
||||
<img src="https://docs.netbird.io/docs-static/img/architecture/high-level-dia.png" width="700"/>
|
||||
</p>
|
||||
|
||||
See a complete [architecture overview](https://docs.netbird.io/about-netbird/how-netbird-works#architecture) for details.
|
||||
|
||||
@@ -4,7 +4,7 @@
|
||||
# sudo podman build -t localhost/netbird:latest -f client/Dockerfile --ignorefile .dockerignore-client .
|
||||
# sudo podman run --rm -it --cap-add={BPF,NET_ADMIN,NET_RAW} localhost/netbird:latest
|
||||
|
||||
FROM alpine:3.23.2
|
||||
FROM alpine:3.22.2
|
||||
# iproute2: busybox doesn't display ip rules properly
|
||||
RUN apk add --no-cache \
|
||||
bash \
|
||||
|
||||
@@ -59,6 +59,7 @@ func init() {
|
||||
|
||||
// Client struct manage the life circle of background service
|
||||
type Client struct {
|
||||
cfgFile string
|
||||
tunAdapter device.TunAdapter
|
||||
iFaceDiscover IFaceDiscover
|
||||
recorder *peer.Status
|
||||
@@ -67,16 +68,18 @@ type Client struct {
|
||||
deviceName string
|
||||
uiVersion string
|
||||
networkChangeListener listener.NetworkChangeListener
|
||||
stateFile string
|
||||
|
||||
connectClient *internal.ConnectClient
|
||||
}
|
||||
|
||||
// NewClient instantiate a new Client
|
||||
func NewClient(androidSDKVersion int, deviceName string, uiVersion string, tunAdapter TunAdapter, iFaceDiscover IFaceDiscover, networkChangeListener NetworkChangeListener) *Client {
|
||||
func NewClient(platformFiles PlatformFiles, androidSDKVersion int, deviceName string, uiVersion string, tunAdapter TunAdapter, iFaceDiscover IFaceDiscover, networkChangeListener NetworkChangeListener) *Client {
|
||||
execWorkaround(androidSDKVersion)
|
||||
|
||||
net.SetAndroidProtectSocketFn(tunAdapter.ProtectSocket)
|
||||
return &Client{
|
||||
cfgFile: platformFiles.ConfigurationFilePath(),
|
||||
deviceName: deviceName,
|
||||
uiVersion: uiVersion,
|
||||
tunAdapter: tunAdapter,
|
||||
@@ -84,20 +87,15 @@ func NewClient(androidSDKVersion int, deviceName string, uiVersion string, tunAd
|
||||
recorder: peer.NewRecorder(""),
|
||||
ctxCancelLock: &sync.Mutex{},
|
||||
networkChangeListener: networkChangeListener,
|
||||
stateFile: platformFiles.StateFilePath(),
|
||||
}
|
||||
}
|
||||
|
||||
// Run start the internal client. It is a blocker function
|
||||
func (c *Client) Run(platformFiles PlatformFiles, urlOpener URLOpener, isAndroidTV bool, dns *DNSList, dnsReadyListener DnsReadyListener, envList *EnvList) error {
|
||||
func (c *Client) Run(urlOpener URLOpener, isAndroidTV bool, dns *DNSList, dnsReadyListener DnsReadyListener, envList *EnvList) error {
|
||||
exportEnvList(envList)
|
||||
|
||||
cfgFile := platformFiles.ConfigurationFilePath()
|
||||
stateFile := platformFiles.StateFilePath()
|
||||
|
||||
log.Infof("Starting client with config: %s, state: %s", cfgFile, stateFile)
|
||||
|
||||
cfg, err := profilemanager.UpdateOrCreateConfig(profilemanager.ConfigInput{
|
||||
ConfigPath: cfgFile,
|
||||
ConfigPath: c.cfgFile,
|
||||
})
|
||||
if err != nil {
|
||||
return err
|
||||
@@ -124,22 +122,16 @@ func (c *Client) Run(platformFiles PlatformFiles, urlOpener URLOpener, isAndroid
|
||||
|
||||
// todo do not throw error in case of cancelled context
|
||||
ctx = internal.CtxInitState(ctx)
|
||||
c.connectClient = internal.NewConnectClient(ctx, cfg, c.recorder, false)
|
||||
return c.connectClient.RunOnAndroid(c.tunAdapter, c.iFaceDiscover, c.networkChangeListener, slices.Clone(dns.items), dnsReadyListener, stateFile)
|
||||
c.connectClient = internal.NewConnectClient(ctx, cfg, c.recorder)
|
||||
return c.connectClient.RunOnAndroid(c.tunAdapter, c.iFaceDiscover, c.networkChangeListener, slices.Clone(dns.items), dnsReadyListener, c.stateFile)
|
||||
}
|
||||
|
||||
// RunWithoutLogin we apply this type of run function when the backed has been started without UI (i.e. after reboot).
|
||||
// In this case make no sense handle registration steps.
|
||||
func (c *Client) RunWithoutLogin(platformFiles PlatformFiles, dns *DNSList, dnsReadyListener DnsReadyListener, envList *EnvList) error {
|
||||
func (c *Client) RunWithoutLogin(dns *DNSList, dnsReadyListener DnsReadyListener, envList *EnvList) error {
|
||||
exportEnvList(envList)
|
||||
|
||||
cfgFile := platformFiles.ConfigurationFilePath()
|
||||
stateFile := platformFiles.StateFilePath()
|
||||
|
||||
log.Infof("Starting client without login with config: %s, state: %s", cfgFile, stateFile)
|
||||
|
||||
cfg, err := profilemanager.UpdateOrCreateConfig(profilemanager.ConfigInput{
|
||||
ConfigPath: cfgFile,
|
||||
ConfigPath: c.cfgFile,
|
||||
})
|
||||
if err != nil {
|
||||
return err
|
||||
@@ -157,8 +149,8 @@ func (c *Client) RunWithoutLogin(platformFiles PlatformFiles, dns *DNSList, dnsR
|
||||
|
||||
// todo do not throw error in case of cancelled context
|
||||
ctx = internal.CtxInitState(ctx)
|
||||
c.connectClient = internal.NewConnectClient(ctx, cfg, c.recorder, false)
|
||||
return c.connectClient.RunOnAndroid(c.tunAdapter, c.iFaceDiscover, c.networkChangeListener, slices.Clone(dns.items), dnsReadyListener, stateFile)
|
||||
c.connectClient = internal.NewConnectClient(ctx, cfg, c.recorder)
|
||||
return c.connectClient.RunOnAndroid(c.tunAdapter, c.iFaceDiscover, c.networkChangeListener, slices.Clone(dns.items), dnsReadyListener, c.stateFile)
|
||||
}
|
||||
|
||||
// Stop the internal client and free the resources
|
||||
|
||||
@@ -1,257 +0,0 @@
|
||||
//go:build android
|
||||
|
||||
package android
|
||||
|
||||
import (
|
||||
"fmt"
|
||||
"os"
|
||||
"path/filepath"
|
||||
"strings"
|
||||
|
||||
log "github.com/sirupsen/logrus"
|
||||
|
||||
"github.com/netbirdio/netbird/client/internal/profilemanager"
|
||||
)
|
||||
|
||||
const (
|
||||
// Android-specific config filename (different from desktop default.json)
|
||||
defaultConfigFilename = "netbird.cfg"
|
||||
// Subdirectory for non-default profiles (must match Java Preferences.java)
|
||||
profilesSubdir = "profiles"
|
||||
// Android uses a single user context per app (non-empty username required by ServiceManager)
|
||||
androidUsername = "android"
|
||||
)
|
||||
|
||||
// Profile represents a profile for gomobile
|
||||
type Profile struct {
|
||||
Name string
|
||||
IsActive bool
|
||||
}
|
||||
|
||||
// ProfileArray wraps profiles for gomobile compatibility
|
||||
type ProfileArray struct {
|
||||
items []*Profile
|
||||
}
|
||||
|
||||
// Length returns the number of profiles
|
||||
func (p *ProfileArray) Length() int {
|
||||
return len(p.items)
|
||||
}
|
||||
|
||||
// Get returns the profile at index i
|
||||
func (p *ProfileArray) Get(i int) *Profile {
|
||||
if i < 0 || i >= len(p.items) {
|
||||
return nil
|
||||
}
|
||||
return p.items[i]
|
||||
}
|
||||
|
||||
/*
|
||||
|
||||
/data/data/io.netbird.client/files/ ← configDir parameter
|
||||
├── netbird.cfg ← Default profile config
|
||||
├── state.json ← Default profile state
|
||||
├── active_profile.json ← Active profile tracker (JSON with Name + Username)
|
||||
└── profiles/ ← Subdirectory for non-default profiles
|
||||
├── work.json ← Work profile config
|
||||
├── work.state.json ← Work profile state
|
||||
├── personal.json ← Personal profile config
|
||||
└── personal.state.json ← Personal profile state
|
||||
*/
|
||||
|
||||
// ProfileManager manages profiles for Android
|
||||
// It wraps the internal profilemanager to provide Android-specific behavior
|
||||
type ProfileManager struct {
|
||||
configDir string
|
||||
serviceMgr *profilemanager.ServiceManager
|
||||
}
|
||||
|
||||
// NewProfileManager creates a new profile manager for Android
|
||||
func NewProfileManager(configDir string) *ProfileManager {
|
||||
// Set the default config path for Android (stored in root configDir, not profiles/)
|
||||
defaultConfigPath := filepath.Join(configDir, defaultConfigFilename)
|
||||
|
||||
// Set global paths for Android
|
||||
profilemanager.DefaultConfigPathDir = configDir
|
||||
profilemanager.DefaultConfigPath = defaultConfigPath
|
||||
profilemanager.ActiveProfileStatePath = filepath.Join(configDir, "active_profile.json")
|
||||
|
||||
// Create ServiceManager with profiles/ subdirectory
|
||||
// This avoids modifying the global ConfigDirOverride for profile listing
|
||||
profilesDir := filepath.Join(configDir, profilesSubdir)
|
||||
serviceMgr := profilemanager.NewServiceManagerWithProfilesDir(defaultConfigPath, profilesDir)
|
||||
|
||||
return &ProfileManager{
|
||||
configDir: configDir,
|
||||
serviceMgr: serviceMgr,
|
||||
}
|
||||
}
|
||||
|
||||
// ListProfiles returns all available profiles
|
||||
func (pm *ProfileManager) ListProfiles() (*ProfileArray, error) {
|
||||
// Use ServiceManager (looks in profiles/ directory, checks active_profile.json for IsActive)
|
||||
internalProfiles, err := pm.serviceMgr.ListProfiles(androidUsername)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("failed to list profiles: %w", err)
|
||||
}
|
||||
|
||||
// Convert internal profiles to Android Profile type
|
||||
var profiles []*Profile
|
||||
for _, p := range internalProfiles {
|
||||
profiles = append(profiles, &Profile{
|
||||
Name: p.Name,
|
||||
IsActive: p.IsActive,
|
||||
})
|
||||
}
|
||||
|
||||
return &ProfileArray{items: profiles}, nil
|
||||
}
|
||||
|
||||
// GetActiveProfile returns the currently active profile name
|
||||
func (pm *ProfileManager) GetActiveProfile() (string, error) {
|
||||
// Use ServiceManager to stay consistent with ListProfiles
|
||||
// ServiceManager uses active_profile.json
|
||||
activeState, err := pm.serviceMgr.GetActiveProfileState()
|
||||
if err != nil {
|
||||
return "", fmt.Errorf("failed to get active profile: %w", err)
|
||||
}
|
||||
return activeState.Name, nil
|
||||
}
|
||||
|
||||
// SwitchProfile switches to a different profile
|
||||
func (pm *ProfileManager) SwitchProfile(profileName string) error {
|
||||
// Use ServiceManager to stay consistent with ListProfiles
|
||||
// ServiceManager uses active_profile.json
|
||||
err := pm.serviceMgr.SetActiveProfileState(&profilemanager.ActiveProfileState{
|
||||
Name: profileName,
|
||||
Username: androidUsername,
|
||||
})
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to switch profile: %w", err)
|
||||
}
|
||||
|
||||
log.Infof("switched to profile: %s", profileName)
|
||||
return nil
|
||||
}
|
||||
|
||||
// AddProfile creates a new profile
|
||||
func (pm *ProfileManager) AddProfile(profileName string) error {
|
||||
// Use ServiceManager (creates profile in profiles/ directory)
|
||||
if err := pm.serviceMgr.AddProfile(profileName, androidUsername); err != nil {
|
||||
return fmt.Errorf("failed to add profile: %w", err)
|
||||
}
|
||||
|
||||
log.Infof("created new profile: %s", profileName)
|
||||
return nil
|
||||
}
|
||||
|
||||
// LogoutProfile logs out from a profile (clears authentication)
|
||||
func (pm *ProfileManager) LogoutProfile(profileName string) error {
|
||||
profileName = sanitizeProfileName(profileName)
|
||||
|
||||
configPath, err := pm.getProfileConfigPath(profileName)
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
|
||||
// Check if profile exists
|
||||
if _, err := os.Stat(configPath); os.IsNotExist(err) {
|
||||
return fmt.Errorf("profile '%s' does not exist", profileName)
|
||||
}
|
||||
|
||||
// Read current config using internal profilemanager
|
||||
config, err := profilemanager.ReadConfig(configPath)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to read profile config: %w", err)
|
||||
}
|
||||
|
||||
// Clear authentication by removing private key and SSH key
|
||||
config.PrivateKey = ""
|
||||
config.SSHKey = ""
|
||||
|
||||
// Save config using internal profilemanager
|
||||
if err := profilemanager.WriteOutConfig(configPath, config); err != nil {
|
||||
return fmt.Errorf("failed to save config: %w", err)
|
||||
}
|
||||
|
||||
log.Infof("logged out from profile: %s", profileName)
|
||||
return nil
|
||||
}
|
||||
|
||||
// RemoveProfile deletes a profile
|
||||
func (pm *ProfileManager) RemoveProfile(profileName string) error {
|
||||
// Use ServiceManager (removes profile from profiles/ directory)
|
||||
if err := pm.serviceMgr.RemoveProfile(profileName, androidUsername); err != nil {
|
||||
return fmt.Errorf("failed to remove profile: %w", err)
|
||||
}
|
||||
|
||||
log.Infof("removed profile: %s", profileName)
|
||||
return nil
|
||||
}
|
||||
|
||||
// getProfileConfigPath returns the config file path for a profile
|
||||
// This is needed for Android-specific path handling (netbird.cfg for default profile)
|
||||
func (pm *ProfileManager) getProfileConfigPath(profileName string) (string, error) {
|
||||
if profileName == "" || profileName == profilemanager.DefaultProfileName {
|
||||
// Android uses netbird.cfg for default profile instead of default.json
|
||||
// Default profile is stored in root configDir, not in profiles/
|
||||
return filepath.Join(pm.configDir, defaultConfigFilename), nil
|
||||
}
|
||||
|
||||
// Non-default profiles are stored in profiles subdirectory
|
||||
// This matches the Java Preferences.java expectation
|
||||
profileName = sanitizeProfileName(profileName)
|
||||
profilesDir := filepath.Join(pm.configDir, profilesSubdir)
|
||||
return filepath.Join(profilesDir, profileName+".json"), nil
|
||||
}
|
||||
|
||||
// GetConfigPath returns the config file path for a given profile
|
||||
// Java should call this instead of constructing paths with Preferences.configFile()
|
||||
func (pm *ProfileManager) GetConfigPath(profileName string) (string, error) {
|
||||
return pm.getProfileConfigPath(profileName)
|
||||
}
|
||||
|
||||
// GetStateFilePath returns the state file path for a given profile
|
||||
// Java should call this instead of constructing paths with Preferences.stateFile()
|
||||
func (pm *ProfileManager) GetStateFilePath(profileName string) (string, error) {
|
||||
if profileName == "" || profileName == profilemanager.DefaultProfileName {
|
||||
return filepath.Join(pm.configDir, "state.json"), nil
|
||||
}
|
||||
|
||||
profileName = sanitizeProfileName(profileName)
|
||||
profilesDir := filepath.Join(pm.configDir, profilesSubdir)
|
||||
return filepath.Join(profilesDir, profileName+".state.json"), nil
|
||||
}
|
||||
|
||||
// GetActiveConfigPath returns the config file path for the currently active profile
|
||||
// Java should call this instead of Preferences.getActiveProfileName() + Preferences.configFile()
|
||||
func (pm *ProfileManager) GetActiveConfigPath() (string, error) {
|
||||
activeProfile, err := pm.GetActiveProfile()
|
||||
if err != nil {
|
||||
return "", fmt.Errorf("failed to get active profile: %w", err)
|
||||
}
|
||||
return pm.GetConfigPath(activeProfile)
|
||||
}
|
||||
|
||||
// GetActiveStateFilePath returns the state file path for the currently active profile
|
||||
// Java should call this instead of Preferences.getActiveProfileName() + Preferences.stateFile()
|
||||
func (pm *ProfileManager) GetActiveStateFilePath() (string, error) {
|
||||
activeProfile, err := pm.GetActiveProfile()
|
||||
if err != nil {
|
||||
return "", fmt.Errorf("failed to get active profile: %w", err)
|
||||
}
|
||||
return pm.GetStateFilePath(activeProfile)
|
||||
}
|
||||
|
||||
// sanitizeProfileName removes invalid characters from profile name
|
||||
func sanitizeProfileName(name string) string {
|
||||
// Keep only alphanumeric, underscore, and hyphen
|
||||
var result strings.Builder
|
||||
for _, r := range name {
|
||||
if (r >= 'a' && r <= 'z') || (r >= 'A' && r <= 'Z') ||
|
||||
(r >= '0' && r <= '9') || r == '_' || r == '-' {
|
||||
result.WriteRune(r)
|
||||
}
|
||||
}
|
||||
return result.String()
|
||||
}
|
||||
@@ -136,7 +136,6 @@ func setLogLevel(cmd *cobra.Command, args []string) error {
|
||||
client := proto.NewDaemonServiceClient(conn)
|
||||
level := server.ParseLogLevel(args[0])
|
||||
if level == proto.LogLevel_UNKNOWN {
|
||||
//nolint
|
||||
return fmt.Errorf("unknown log level: %s. Available levels are: panic, fatal, error, warn, info, debug, trace\n", args[0])
|
||||
}
|
||||
|
||||
@@ -314,8 +313,9 @@ func getStatusOutput(cmd *cobra.Command, anon bool) string {
|
||||
profName = activeProf.Name
|
||||
}
|
||||
|
||||
overview := nbstatus.ConvertToStatusOutputOverview(statusResp, anon, "", nil, nil, nil, "", profName)
|
||||
statusOutputString = overview.FullDetailSummary()
|
||||
statusOutputString = nbstatus.ParseToFullDetailSummary(
|
||||
nbstatus.ConvertToStatusOutputOverview(statusResp, anon, "", nil, nil, nil, "", profName),
|
||||
)
|
||||
}
|
||||
return statusOutputString
|
||||
}
|
||||
|
||||
@@ -81,7 +81,6 @@ var loginCmd = &cobra.Command{
|
||||
func doDaemonLogin(ctx context.Context, cmd *cobra.Command, providedSetupKey string, activeProf *profilemanager.Profile, username string, pm *profilemanager.ProfileManager) error {
|
||||
conn, err := DialClientGRPCServer(ctx, daemonAddr)
|
||||
if err != nil {
|
||||
//nolint
|
||||
return fmt.Errorf("failed to connect to daemon error: %v\n"+
|
||||
"If the daemon is not running please run: "+
|
||||
"\nnetbird service install \nnetbird service start\n", err)
|
||||
@@ -207,7 +206,6 @@ func switchProfileOnDaemon(ctx context.Context, pm *profilemanager.ProfileManage
|
||||
func switchProfile(ctx context.Context, profileName string, username string) error {
|
||||
conn, err := DialClientGRPCServer(ctx, daemonAddr)
|
||||
if err != nil {
|
||||
//nolint
|
||||
return fmt.Errorf("failed to connect to daemon error: %v\n"+
|
||||
"If the daemon is not running please run: "+
|
||||
"\nnetbird service install \nnetbird service start\n", err)
|
||||
|
||||
@@ -1,4 +1,5 @@
|
||||
//go:build pprof
|
||||
// +build pprof
|
||||
|
||||
package cmd
|
||||
|
||||
|
||||
@@ -85,9 +85,6 @@ var (
|
||||
|
||||
// Execute executes the root command.
|
||||
func Execute() error {
|
||||
if isUpdateBinary() {
|
||||
return updateCmd.Execute()
|
||||
}
|
||||
return rootCmd.Execute()
|
||||
}
|
||||
|
||||
@@ -390,7 +387,6 @@ func getClient(cmd *cobra.Command) (*grpc.ClientConn, error) {
|
||||
|
||||
conn, err := DialClientGRPCServer(cmd.Context(), daemonAddr)
|
||||
if err != nil {
|
||||
//nolint
|
||||
return nil, fmt.Errorf("failed to connect to daemon error: %v\n"+
|
||||
"If the daemon is not running please run: "+
|
||||
"\nnetbird service install \nnetbird service start\n", err)
|
||||
|
||||
@@ -1,176 +0,0 @@
|
||||
package main
|
||||
|
||||
import (
|
||||
"fmt"
|
||||
"os"
|
||||
"time"
|
||||
|
||||
"github.com/spf13/cobra"
|
||||
|
||||
"github.com/netbirdio/netbird/client/internal/updatemanager/reposign"
|
||||
)
|
||||
|
||||
var (
|
||||
bundlePubKeysRootPrivKeyFile string
|
||||
bundlePubKeysPubKeyFiles []string
|
||||
bundlePubKeysFile string
|
||||
|
||||
createArtifactKeyRootPrivKeyFile string
|
||||
createArtifactKeyPrivKeyFile string
|
||||
createArtifactKeyPubKeyFile string
|
||||
createArtifactKeyExpiration time.Duration
|
||||
)
|
||||
|
||||
var createArtifactKeyCmd = &cobra.Command{
|
||||
Use: "create-artifact-key",
|
||||
Short: "Create a new artifact signing key",
|
||||
Long: `Generate a new artifact signing key pair signed by the root private key.
|
||||
The artifact key will be used to sign software artifacts/updates.`,
|
||||
SilenceUsage: true,
|
||||
RunE: func(cmd *cobra.Command, args []string) error {
|
||||
if createArtifactKeyExpiration <= 0 {
|
||||
return fmt.Errorf("--expiration must be a positive duration (e.g., 720h, 365d, 8760h)")
|
||||
}
|
||||
|
||||
if err := handleCreateArtifactKey(cmd, createArtifactKeyRootPrivKeyFile, createArtifactKeyPrivKeyFile, createArtifactKeyPubKeyFile, createArtifactKeyExpiration); err != nil {
|
||||
return fmt.Errorf("failed to create artifact key: %w", err)
|
||||
}
|
||||
return nil
|
||||
},
|
||||
}
|
||||
|
||||
var bundlePubKeysCmd = &cobra.Command{
|
||||
Use: "bundle-pub-keys",
|
||||
Short: "Bundle multiple artifact public keys into a signed package",
|
||||
Long: `Bundle one or more artifact public keys into a signed package using the root private key.
|
||||
This command is typically used to distribute or authorize a set of valid artifact signing keys.`,
|
||||
RunE: func(cmd *cobra.Command, args []string) error {
|
||||
if len(bundlePubKeysPubKeyFiles) == 0 {
|
||||
return fmt.Errorf("at least one --artifact-pub-key-file must be provided")
|
||||
}
|
||||
|
||||
if err := handleBundlePubKeys(cmd, bundlePubKeysRootPrivKeyFile, bundlePubKeysPubKeyFiles, bundlePubKeysFile); err != nil {
|
||||
return fmt.Errorf("failed to bundle public keys: %w", err)
|
||||
}
|
||||
return nil
|
||||
},
|
||||
}
|
||||
|
||||
func init() {
|
||||
rootCmd.AddCommand(createArtifactKeyCmd)
|
||||
|
||||
createArtifactKeyCmd.Flags().StringVar(&createArtifactKeyRootPrivKeyFile, "root-private-key-file", "", "Path to the root private key file used to sign the artifact key")
|
||||
createArtifactKeyCmd.Flags().StringVar(&createArtifactKeyPrivKeyFile, "artifact-priv-key-file", "", "Path where the artifact private key will be saved")
|
||||
createArtifactKeyCmd.Flags().StringVar(&createArtifactKeyPubKeyFile, "artifact-pub-key-file", "", "Path where the artifact public key will be saved")
|
||||
createArtifactKeyCmd.Flags().DurationVar(&createArtifactKeyExpiration, "expiration", 0, "Expiration duration for the artifact key (e.g., 720h, 365d, 8760h)")
|
||||
|
||||
if err := createArtifactKeyCmd.MarkFlagRequired("root-private-key-file"); err != nil {
|
||||
panic(fmt.Errorf("mark root-private-key-file as required: %w", err))
|
||||
}
|
||||
if err := createArtifactKeyCmd.MarkFlagRequired("artifact-priv-key-file"); err != nil {
|
||||
panic(fmt.Errorf("mark artifact-priv-key-file as required: %w", err))
|
||||
}
|
||||
if err := createArtifactKeyCmd.MarkFlagRequired("artifact-pub-key-file"); err != nil {
|
||||
panic(fmt.Errorf("mark artifact-pub-key-file as required: %w", err))
|
||||
}
|
||||
if err := createArtifactKeyCmd.MarkFlagRequired("expiration"); err != nil {
|
||||
panic(fmt.Errorf("mark expiration as required: %w", err))
|
||||
}
|
||||
|
||||
rootCmd.AddCommand(bundlePubKeysCmd)
|
||||
|
||||
bundlePubKeysCmd.Flags().StringVar(&bundlePubKeysRootPrivKeyFile, "root-private-key-file", "", "Path to the root private key file used to sign the bundle")
|
||||
bundlePubKeysCmd.Flags().StringArrayVar(&bundlePubKeysPubKeyFiles, "artifact-pub-key-file", nil, "Path(s) to the artifact public key files to include in the bundle (can be repeated)")
|
||||
bundlePubKeysCmd.Flags().StringVar(&bundlePubKeysFile, "bundle-pub-key-file", "", "Path where the public keys will be saved")
|
||||
|
||||
if err := bundlePubKeysCmd.MarkFlagRequired("root-private-key-file"); err != nil {
|
||||
panic(fmt.Errorf("mark root-private-key-file as required: %w", err))
|
||||
}
|
||||
if err := bundlePubKeysCmd.MarkFlagRequired("artifact-pub-key-file"); err != nil {
|
||||
panic(fmt.Errorf("mark artifact-pub-key-file as required: %w", err))
|
||||
}
|
||||
if err := bundlePubKeysCmd.MarkFlagRequired("bundle-pub-key-file"); err != nil {
|
||||
panic(fmt.Errorf("mark bundle-pub-key-file as required: %w", err))
|
||||
}
|
||||
}
|
||||
|
||||
func handleCreateArtifactKey(cmd *cobra.Command, rootPrivKeyFile, artifactPrivKeyFile, artifactPubKeyFile string, expiration time.Duration) error {
|
||||
cmd.Println("Creating new artifact signing key...")
|
||||
|
||||
privKeyPEM, err := os.ReadFile(rootPrivKeyFile)
|
||||
if err != nil {
|
||||
return fmt.Errorf("read root private key file: %w", err)
|
||||
}
|
||||
|
||||
privateRootKey, err := reposign.ParseRootKey(privKeyPEM)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to parse private root key: %w", err)
|
||||
}
|
||||
|
||||
artifactKey, privPEM, pubPEM, signature, err := reposign.GenerateArtifactKey(privateRootKey, expiration)
|
||||
if err != nil {
|
||||
return fmt.Errorf("generate artifact key: %w", err)
|
||||
}
|
||||
|
||||
if err := os.WriteFile(artifactPrivKeyFile, privPEM, 0o600); err != nil {
|
||||
return fmt.Errorf("write private key file (%s): %w", artifactPrivKeyFile, err)
|
||||
}
|
||||
|
||||
if err := os.WriteFile(artifactPubKeyFile, pubPEM, 0o600); err != nil {
|
||||
return fmt.Errorf("write public key file (%s): %w", artifactPubKeyFile, err)
|
||||
}
|
||||
|
||||
signatureFile := artifactPubKeyFile + ".sig"
|
||||
if err := os.WriteFile(signatureFile, signature, 0o600); err != nil {
|
||||
return fmt.Errorf("write signature file (%s): %w", signatureFile, err)
|
||||
}
|
||||
|
||||
cmd.Printf("✅ Artifact key created successfully.\n")
|
||||
cmd.Printf("%s\n", artifactKey.String())
|
||||
return nil
|
||||
}
|
||||
|
||||
func handleBundlePubKeys(cmd *cobra.Command, rootPrivKeyFile string, artifactPubKeyFiles []string, bundlePubKeysFile string) error {
|
||||
cmd.Println("📦 Bundling public keys into signed package...")
|
||||
|
||||
privKeyPEM, err := os.ReadFile(rootPrivKeyFile)
|
||||
if err != nil {
|
||||
return fmt.Errorf("read root private key file: %w", err)
|
||||
}
|
||||
|
||||
privateRootKey, err := reposign.ParseRootKey(privKeyPEM)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to parse private root key: %w", err)
|
||||
}
|
||||
|
||||
publicKeys := make([]reposign.PublicKey, 0, len(artifactPubKeyFiles))
|
||||
for _, pubFile := range artifactPubKeyFiles {
|
||||
pubPem, err := os.ReadFile(pubFile)
|
||||
if err != nil {
|
||||
return fmt.Errorf("read public key file: %w", err)
|
||||
}
|
||||
|
||||
pk, err := reposign.ParseArtifactPubKey(pubPem)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to parse artifact key: %w", err)
|
||||
}
|
||||
publicKeys = append(publicKeys, pk)
|
||||
}
|
||||
|
||||
parsedKeys, signature, err := reposign.BundleArtifactKeys(privateRootKey, publicKeys)
|
||||
if err != nil {
|
||||
return fmt.Errorf("bundle artifact keys: %w", err)
|
||||
}
|
||||
|
||||
if err := os.WriteFile(bundlePubKeysFile, parsedKeys, 0o600); err != nil {
|
||||
return fmt.Errorf("write public keys file (%s): %w", bundlePubKeysFile, err)
|
||||
}
|
||||
|
||||
signatureFile := bundlePubKeysFile + ".sig"
|
||||
if err := os.WriteFile(signatureFile, signature, 0o600); err != nil {
|
||||
return fmt.Errorf("write signature file (%s): %w", signatureFile, err)
|
||||
}
|
||||
|
||||
cmd.Printf("✅ Bundle created with %d public keys.\n", len(artifactPubKeyFiles))
|
||||
return nil
|
||||
}
|
||||
@@ -1,276 +0,0 @@
|
||||
package main
|
||||
|
||||
import (
|
||||
"fmt"
|
||||
"os"
|
||||
|
||||
"github.com/spf13/cobra"
|
||||
|
||||
"github.com/netbirdio/netbird/client/internal/updatemanager/reposign"
|
||||
)
|
||||
|
||||
const (
|
||||
envArtifactPrivateKey = "NB_ARTIFACT_PRIV_KEY"
|
||||
)
|
||||
|
||||
var (
|
||||
signArtifactPrivKeyFile string
|
||||
signArtifactArtifactFile string
|
||||
|
||||
verifyArtifactPubKeyFile string
|
||||
verifyArtifactFile string
|
||||
verifyArtifactSignatureFile string
|
||||
|
||||
verifyArtifactKeyPubKeyFile string
|
||||
verifyArtifactKeyRootPubKeyFile string
|
||||
verifyArtifactKeySignatureFile string
|
||||
verifyArtifactKeyRevocationFile string
|
||||
)
|
||||
|
||||
var signArtifactCmd = &cobra.Command{
|
||||
Use: "sign-artifact",
|
||||
Short: "Sign an artifact using an artifact private key",
|
||||
Long: `Sign a software artifact (e.g., update bundle or binary) using the artifact's private key.
|
||||
This command produces a detached signature that can be verified using the corresponding artifact public key.`,
|
||||
SilenceUsage: true,
|
||||
RunE: func(cmd *cobra.Command, args []string) error {
|
||||
if err := handleSignArtifact(cmd, signArtifactPrivKeyFile, signArtifactArtifactFile); err != nil {
|
||||
return fmt.Errorf("failed to sign artifact: %w", err)
|
||||
}
|
||||
return nil
|
||||
},
|
||||
}
|
||||
|
||||
var verifyArtifactCmd = &cobra.Command{
|
||||
Use: "verify-artifact",
|
||||
Short: "Verify an artifact signature using an artifact public key",
|
||||
Long: `Verify a software artifact signature using the artifact's public key.`,
|
||||
SilenceUsage: true,
|
||||
RunE: func(cmd *cobra.Command, args []string) error {
|
||||
if err := handleVerifyArtifact(cmd, verifyArtifactPubKeyFile, verifyArtifactFile, verifyArtifactSignatureFile); err != nil {
|
||||
return fmt.Errorf("failed to verify artifact: %w", err)
|
||||
}
|
||||
return nil
|
||||
},
|
||||
}
|
||||
|
||||
var verifyArtifactKeyCmd = &cobra.Command{
|
||||
Use: "verify-artifact-key",
|
||||
Short: "Verify an artifact public key was signed by a root key",
|
||||
Long: `Verify that an artifact public key (or bundle) was properly signed by a root key.
|
||||
This validates the chain of trust from the root key to the artifact key.`,
|
||||
SilenceUsage: true,
|
||||
RunE: func(cmd *cobra.Command, args []string) error {
|
||||
if err := handleVerifyArtifactKey(cmd, verifyArtifactKeyPubKeyFile, verifyArtifactKeyRootPubKeyFile, verifyArtifactKeySignatureFile, verifyArtifactKeyRevocationFile); err != nil {
|
||||
return fmt.Errorf("failed to verify artifact key: %w", err)
|
||||
}
|
||||
return nil
|
||||
},
|
||||
}
|
||||
|
||||
func init() {
|
||||
rootCmd.AddCommand(signArtifactCmd)
|
||||
rootCmd.AddCommand(verifyArtifactCmd)
|
||||
rootCmd.AddCommand(verifyArtifactKeyCmd)
|
||||
|
||||
signArtifactCmd.Flags().StringVar(&signArtifactPrivKeyFile, "artifact-key-file", "", fmt.Sprintf("Path to the artifact private key file used for signing (or set %s env var)", envArtifactPrivateKey))
|
||||
signArtifactCmd.Flags().StringVar(&signArtifactArtifactFile, "artifact-file", "", "Path to the artifact to be signed")
|
||||
|
||||
// artifact-file is required, but artifact-key-file can come from env var
|
||||
if err := signArtifactCmd.MarkFlagRequired("artifact-file"); err != nil {
|
||||
panic(fmt.Errorf("mark artifact-file as required: %w", err))
|
||||
}
|
||||
|
||||
verifyArtifactCmd.Flags().StringVar(&verifyArtifactPubKeyFile, "artifact-public-key-file", "", "Path to the artifact public key file")
|
||||
verifyArtifactCmd.Flags().StringVar(&verifyArtifactFile, "artifact-file", "", "Path to the artifact to be verified")
|
||||
verifyArtifactCmd.Flags().StringVar(&verifyArtifactSignatureFile, "signature-file", "", "Path to the signature file")
|
||||
|
||||
if err := verifyArtifactCmd.MarkFlagRequired("artifact-public-key-file"); err != nil {
|
||||
panic(fmt.Errorf("mark artifact-public-key-file as required: %w", err))
|
||||
}
|
||||
if err := verifyArtifactCmd.MarkFlagRequired("artifact-file"); err != nil {
|
||||
panic(fmt.Errorf("mark artifact-file as required: %w", err))
|
||||
}
|
||||
if err := verifyArtifactCmd.MarkFlagRequired("signature-file"); err != nil {
|
||||
panic(fmt.Errorf("mark signature-file as required: %w", err))
|
||||
}
|
||||
|
||||
verifyArtifactKeyCmd.Flags().StringVar(&verifyArtifactKeyPubKeyFile, "artifact-key-file", "", "Path to the artifact public key file or bundle")
|
||||
verifyArtifactKeyCmd.Flags().StringVar(&verifyArtifactKeyRootPubKeyFile, "root-key-file", "", "Path to the root public key file or bundle")
|
||||
verifyArtifactKeyCmd.Flags().StringVar(&verifyArtifactKeySignatureFile, "signature-file", "", "Path to the signature file")
|
||||
verifyArtifactKeyCmd.Flags().StringVar(&verifyArtifactKeyRevocationFile, "revocation-file", "", "Path to the revocation list file (optional)")
|
||||
|
||||
if err := verifyArtifactKeyCmd.MarkFlagRequired("artifact-key-file"); err != nil {
|
||||
panic(fmt.Errorf("mark artifact-key-file as required: %w", err))
|
||||
}
|
||||
if err := verifyArtifactKeyCmd.MarkFlagRequired("root-key-file"); err != nil {
|
||||
panic(fmt.Errorf("mark root-key-file as required: %w", err))
|
||||
}
|
||||
if err := verifyArtifactKeyCmd.MarkFlagRequired("signature-file"); err != nil {
|
||||
panic(fmt.Errorf("mark signature-file as required: %w", err))
|
||||
}
|
||||
}
|
||||
|
||||
func handleSignArtifact(cmd *cobra.Command, privKeyFile, artifactFile string) error {
|
||||
cmd.Println("🖋️ Signing artifact...")
|
||||
|
||||
// Load private key from env var or file
|
||||
var privKeyPEM []byte
|
||||
var err error
|
||||
|
||||
if envKey := os.Getenv(envArtifactPrivateKey); envKey != "" {
|
||||
// Use key from environment variable
|
||||
privKeyPEM = []byte(envKey)
|
||||
} else if privKeyFile != "" {
|
||||
// Fall back to file
|
||||
privKeyPEM, err = os.ReadFile(privKeyFile)
|
||||
if err != nil {
|
||||
return fmt.Errorf("read private key file: %w", err)
|
||||
}
|
||||
} else {
|
||||
return fmt.Errorf("artifact private key must be provided via %s environment variable or --artifact-key-file flag", envArtifactPrivateKey)
|
||||
}
|
||||
|
||||
privateKey, err := reposign.ParseArtifactKey(privKeyPEM)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to parse artifact private key: %w", err)
|
||||
}
|
||||
|
||||
artifactData, err := os.ReadFile(artifactFile)
|
||||
if err != nil {
|
||||
return fmt.Errorf("read artifact file: %w", err)
|
||||
}
|
||||
|
||||
signature, err := reposign.SignData(privateKey, artifactData)
|
||||
if err != nil {
|
||||
return fmt.Errorf("sign artifact: %w", err)
|
||||
}
|
||||
|
||||
sigFile := artifactFile + ".sig"
|
||||
if err := os.WriteFile(artifactFile+".sig", signature, 0o600); err != nil {
|
||||
return fmt.Errorf("write signature file (%s): %w", sigFile, err)
|
||||
}
|
||||
|
||||
cmd.Printf("✅ Artifact signed successfully.\n")
|
||||
cmd.Printf("Signature file: %s\n", sigFile)
|
||||
return nil
|
||||
}
|
||||
|
||||
func handleVerifyArtifact(cmd *cobra.Command, pubKeyFile, artifactFile, signatureFile string) error {
|
||||
cmd.Println("🔍 Verifying artifact...")
|
||||
|
||||
// Read artifact public key
|
||||
pubKeyPEM, err := os.ReadFile(pubKeyFile)
|
||||
if err != nil {
|
||||
return fmt.Errorf("read public key file: %w", err)
|
||||
}
|
||||
|
||||
publicKey, err := reposign.ParseArtifactPubKey(pubKeyPEM)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to parse artifact public key: %w", err)
|
||||
}
|
||||
|
||||
// Read artifact data
|
||||
artifactData, err := os.ReadFile(artifactFile)
|
||||
if err != nil {
|
||||
return fmt.Errorf("read artifact file: %w", err)
|
||||
}
|
||||
|
||||
// Read signature
|
||||
sigBytes, err := os.ReadFile(signatureFile)
|
||||
if err != nil {
|
||||
return fmt.Errorf("read signature file: %w", err)
|
||||
}
|
||||
|
||||
signature, err := reposign.ParseSignature(sigBytes)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to parse signature: %w", err)
|
||||
}
|
||||
|
||||
// Validate artifact
|
||||
if err := reposign.ValidateArtifact([]reposign.PublicKey{publicKey}, artifactData, *signature); err != nil {
|
||||
return fmt.Errorf("artifact verification failed: %w", err)
|
||||
}
|
||||
|
||||
cmd.Println("✅ Artifact signature is valid")
|
||||
cmd.Printf("Artifact: %s\n", artifactFile)
|
||||
cmd.Printf("Signed by key: %s\n", signature.KeyID)
|
||||
cmd.Printf("Signature timestamp: %s\n", signature.Timestamp.Format("2006-01-02 15:04:05 MST"))
|
||||
return nil
|
||||
}
|
||||
|
||||
func handleVerifyArtifactKey(cmd *cobra.Command, artifactKeyFile, rootKeyFile, signatureFile, revocationFile string) error {
|
||||
cmd.Println("🔍 Verifying artifact key...")
|
||||
|
||||
// Read artifact key data
|
||||
artifactKeyData, err := os.ReadFile(artifactKeyFile)
|
||||
if err != nil {
|
||||
return fmt.Errorf("read artifact key file: %w", err)
|
||||
}
|
||||
|
||||
// Read root public key(s)
|
||||
rootKeyData, err := os.ReadFile(rootKeyFile)
|
||||
if err != nil {
|
||||
return fmt.Errorf("read root key file: %w", err)
|
||||
}
|
||||
|
||||
rootPublicKeys, err := parseRootPublicKeys(rootKeyData)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to parse root public key(s): %w", err)
|
||||
}
|
||||
|
||||
// Read signature
|
||||
sigBytes, err := os.ReadFile(signatureFile)
|
||||
if err != nil {
|
||||
return fmt.Errorf("read signature file: %w", err)
|
||||
}
|
||||
|
||||
signature, err := reposign.ParseSignature(sigBytes)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to parse signature: %w", err)
|
||||
}
|
||||
|
||||
// Read optional revocation list
|
||||
var revocationList *reposign.RevocationList
|
||||
if revocationFile != "" {
|
||||
revData, err := os.ReadFile(revocationFile)
|
||||
if err != nil {
|
||||
return fmt.Errorf("read revocation file: %w", err)
|
||||
}
|
||||
|
||||
revocationList, err = reposign.ParseRevocationList(revData)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to parse revocation list: %w", err)
|
||||
}
|
||||
}
|
||||
|
||||
// Validate artifact key(s)
|
||||
validKeys, err := reposign.ValidateArtifactKeys(rootPublicKeys, artifactKeyData, *signature, revocationList)
|
||||
if err != nil {
|
||||
return fmt.Errorf("artifact key verification failed: %w", err)
|
||||
}
|
||||
|
||||
cmd.Println("✅ Artifact key(s) verified successfully")
|
||||
cmd.Printf("Signed by root key: %s\n", signature.KeyID)
|
||||
cmd.Printf("Signature timestamp: %s\n", signature.Timestamp.Format("2006-01-02 15:04:05 MST"))
|
||||
cmd.Printf("\nValid artifact keys (%d):\n", len(validKeys))
|
||||
for i, key := range validKeys {
|
||||
cmd.Printf(" [%d] Key ID: %s\n", i+1, key.Metadata.ID)
|
||||
cmd.Printf(" Created: %s\n", key.Metadata.CreatedAt.Format("2006-01-02 15:04:05 MST"))
|
||||
if !key.Metadata.ExpiresAt.IsZero() {
|
||||
cmd.Printf(" Expires: %s\n", key.Metadata.ExpiresAt.Format("2006-01-02 15:04:05 MST"))
|
||||
} else {
|
||||
cmd.Printf(" Expires: Never\n")
|
||||
}
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
// parseRootPublicKeys parses a root public key from PEM data
|
||||
func parseRootPublicKeys(data []byte) ([]reposign.PublicKey, error) {
|
||||
key, err := reposign.ParseRootPublicKey(data)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
return []reposign.PublicKey{key}, nil
|
||||
}
|
||||
@@ -1,21 +0,0 @@
|
||||
package main
|
||||
|
||||
import (
|
||||
"os"
|
||||
|
||||
"github.com/spf13/cobra"
|
||||
)
|
||||
|
||||
var rootCmd = &cobra.Command{
|
||||
Use: "signer",
|
||||
Short: "A CLI tool for managing cryptographic keys and artifacts",
|
||||
Long: `signer is a command-line tool that helps you manage
|
||||
root keys, artifact keys, and revocation lists securely.`,
|
||||
}
|
||||
|
||||
func main() {
|
||||
if err := rootCmd.Execute(); err != nil {
|
||||
rootCmd.Println(err)
|
||||
os.Exit(1)
|
||||
}
|
||||
}
|
||||
@@ -1,220 +0,0 @@
|
||||
package main
|
||||
|
||||
import (
|
||||
"fmt"
|
||||
"os"
|
||||
"time"
|
||||
|
||||
"github.com/spf13/cobra"
|
||||
|
||||
"github.com/netbirdio/netbird/client/internal/updatemanager/reposign"
|
||||
)
|
||||
|
||||
const (
|
||||
defaultRevocationListExpiration = 365 * 24 * time.Hour // 1 year
|
||||
)
|
||||
|
||||
var (
|
||||
keyID string
|
||||
revocationListFile string
|
||||
privateRootKeyFile string
|
||||
publicRootKeyFile string
|
||||
signatureFile string
|
||||
expirationDuration time.Duration
|
||||
)
|
||||
|
||||
var createRevocationListCmd = &cobra.Command{
|
||||
Use: "create-revocation-list",
|
||||
Short: "Create a new revocation list signed by the private root key",
|
||||
SilenceUsage: true,
|
||||
RunE: func(cmd *cobra.Command, args []string) error {
|
||||
return handleCreateRevocationList(cmd, revocationListFile, privateRootKeyFile)
|
||||
},
|
||||
}
|
||||
|
||||
var extendRevocationListCmd = &cobra.Command{
|
||||
Use: "extend-revocation-list",
|
||||
Short: "Extend an existing revocation list with a given key ID",
|
||||
SilenceUsage: true,
|
||||
RunE: func(cmd *cobra.Command, args []string) error {
|
||||
return handleExtendRevocationList(cmd, keyID, revocationListFile, privateRootKeyFile)
|
||||
},
|
||||
}
|
||||
|
||||
var verifyRevocationListCmd = &cobra.Command{
|
||||
Use: "verify-revocation-list",
|
||||
Short: "Verify a revocation list signature using the public root key",
|
||||
SilenceUsage: true,
|
||||
RunE: func(cmd *cobra.Command, args []string) error {
|
||||
return handleVerifyRevocationList(cmd, revocationListFile, signatureFile, publicRootKeyFile)
|
||||
},
|
||||
}
|
||||
|
||||
func init() {
|
||||
rootCmd.AddCommand(createRevocationListCmd)
|
||||
rootCmd.AddCommand(extendRevocationListCmd)
|
||||
rootCmd.AddCommand(verifyRevocationListCmd)
|
||||
|
||||
createRevocationListCmd.Flags().StringVar(&revocationListFile, "revocation-list-file", "", "Path to the existing revocation list file")
|
||||
createRevocationListCmd.Flags().StringVar(&privateRootKeyFile, "private-root-key", "", "Path to the private root key PEM file")
|
||||
createRevocationListCmd.Flags().DurationVar(&expirationDuration, "expiration", defaultRevocationListExpiration, "Expiration duration for the revocation list (e.g., 8760h for 1 year)")
|
||||
if err := createRevocationListCmd.MarkFlagRequired("revocation-list-file"); err != nil {
|
||||
panic(err)
|
||||
}
|
||||
if err := createRevocationListCmd.MarkFlagRequired("private-root-key"); err != nil {
|
||||
panic(err)
|
||||
}
|
||||
|
||||
extendRevocationListCmd.Flags().StringVar(&keyID, "key-id", "", "ID of the key to extend the revocation list for")
|
||||
extendRevocationListCmd.Flags().StringVar(&revocationListFile, "revocation-list-file", "", "Path to the existing revocation list file")
|
||||
extendRevocationListCmd.Flags().StringVar(&privateRootKeyFile, "private-root-key", "", "Path to the private root key PEM file")
|
||||
extendRevocationListCmd.Flags().DurationVar(&expirationDuration, "expiration", defaultRevocationListExpiration, "Expiration duration for the revocation list (e.g., 8760h for 1 year)")
|
||||
if err := extendRevocationListCmd.MarkFlagRequired("key-id"); err != nil {
|
||||
panic(err)
|
||||
}
|
||||
if err := extendRevocationListCmd.MarkFlagRequired("revocation-list-file"); err != nil {
|
||||
panic(err)
|
||||
}
|
||||
if err := extendRevocationListCmd.MarkFlagRequired("private-root-key"); err != nil {
|
||||
panic(err)
|
||||
}
|
||||
|
||||
verifyRevocationListCmd.Flags().StringVar(&revocationListFile, "revocation-list-file", "", "Path to the revocation list file")
|
||||
verifyRevocationListCmd.Flags().StringVar(&signatureFile, "signature-file", "", "Path to the signature file")
|
||||
verifyRevocationListCmd.Flags().StringVar(&publicRootKeyFile, "public-root-key", "", "Path to the public root key PEM file")
|
||||
if err := verifyRevocationListCmd.MarkFlagRequired("revocation-list-file"); err != nil {
|
||||
panic(err)
|
||||
}
|
||||
if err := verifyRevocationListCmd.MarkFlagRequired("signature-file"); err != nil {
|
||||
panic(err)
|
||||
}
|
||||
if err := verifyRevocationListCmd.MarkFlagRequired("public-root-key"); err != nil {
|
||||
panic(err)
|
||||
}
|
||||
}
|
||||
|
||||
func handleCreateRevocationList(cmd *cobra.Command, revocationListFile string, privateRootKeyFile string) error {
|
||||
privKeyPEM, err := os.ReadFile(privateRootKeyFile)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to read private root key file: %w", err)
|
||||
}
|
||||
|
||||
privateRootKey, err := reposign.ParseRootKey(privKeyPEM)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to parse private root key: %w", err)
|
||||
}
|
||||
|
||||
rlBytes, sigBytes, err := reposign.CreateRevocationList(*privateRootKey, expirationDuration)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to create revocation list: %w", err)
|
||||
}
|
||||
|
||||
if err := writeOutputFiles(revocationListFile, revocationListFile+".sig", rlBytes, sigBytes); err != nil {
|
||||
return fmt.Errorf("failed to write output files: %w", err)
|
||||
}
|
||||
|
||||
cmd.Println("✅ Revocation list created successfully")
|
||||
return nil
|
||||
}
|
||||
|
||||
func handleExtendRevocationList(cmd *cobra.Command, keyID, revocationListFile, privateRootKeyFile string) error {
|
||||
privKeyPEM, err := os.ReadFile(privateRootKeyFile)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to read private root key file: %w", err)
|
||||
}
|
||||
|
||||
privateRootKey, err := reposign.ParseRootKey(privKeyPEM)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to parse private root key: %w", err)
|
||||
}
|
||||
|
||||
rlBytes, err := os.ReadFile(revocationListFile)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to read revocation list file: %w", err)
|
||||
}
|
||||
|
||||
rl, err := reposign.ParseRevocationList(rlBytes)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to parse revocation list: %w", err)
|
||||
}
|
||||
|
||||
kid, err := reposign.ParseKeyID(keyID)
|
||||
if err != nil {
|
||||
return fmt.Errorf("invalid key ID: %w", err)
|
||||
}
|
||||
|
||||
newRLBytes, sigBytes, err := reposign.ExtendRevocationList(*privateRootKey, *rl, kid, expirationDuration)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to extend revocation list: %w", err)
|
||||
}
|
||||
|
||||
if err := writeOutputFiles(revocationListFile, revocationListFile+".sig", newRLBytes, sigBytes); err != nil {
|
||||
return fmt.Errorf("failed to write output files: %w", err)
|
||||
}
|
||||
|
||||
cmd.Println("✅ Revocation list extended successfully")
|
||||
return nil
|
||||
}
|
||||
|
||||
func handleVerifyRevocationList(cmd *cobra.Command, revocationListFile, signatureFile, publicRootKeyFile string) error {
|
||||
// Read revocation list file
|
||||
rlBytes, err := os.ReadFile(revocationListFile)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to read revocation list file: %w", err)
|
||||
}
|
||||
|
||||
// Read signature file
|
||||
sigBytes, err := os.ReadFile(signatureFile)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to read signature file: %w", err)
|
||||
}
|
||||
|
||||
// Read public root key file
|
||||
pubKeyPEM, err := os.ReadFile(publicRootKeyFile)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to read public root key file: %w", err)
|
||||
}
|
||||
|
||||
// Parse public root key
|
||||
publicKey, err := reposign.ParseRootPublicKey(pubKeyPEM)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to parse public root key: %w", err)
|
||||
}
|
||||
|
||||
// Parse signature
|
||||
signature, err := reposign.ParseSignature(sigBytes)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to parse signature: %w", err)
|
||||
}
|
||||
|
||||
// Validate revocation list
|
||||
rl, err := reposign.ValidateRevocationList([]reposign.PublicKey{publicKey}, rlBytes, *signature)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to validate revocation list: %w", err)
|
||||
}
|
||||
|
||||
// Display results
|
||||
cmd.Println("✅ Revocation list signature is valid")
|
||||
cmd.Printf("Last Updated: %s\n", rl.LastUpdated.Format(time.RFC3339))
|
||||
cmd.Printf("Expires At: %s\n", rl.ExpiresAt.Format(time.RFC3339))
|
||||
cmd.Printf("Number of revoked keys: %d\n", len(rl.Revoked))
|
||||
|
||||
if len(rl.Revoked) > 0 {
|
||||
cmd.Println("\nRevoked Keys:")
|
||||
for keyID, revokedTime := range rl.Revoked {
|
||||
cmd.Printf(" - %s (revoked at: %s)\n", keyID, revokedTime.Format(time.RFC3339))
|
||||
}
|
||||
}
|
||||
|
||||
return nil
|
||||
}
|
||||
|
||||
func writeOutputFiles(rlPath, sigPath string, rlBytes, sigBytes []byte) error {
|
||||
if err := os.WriteFile(rlPath, rlBytes, 0o600); err != nil {
|
||||
return fmt.Errorf("failed to write revocation list file: %w", err)
|
||||
}
|
||||
if err := os.WriteFile(sigPath, sigBytes, 0o600); err != nil {
|
||||
return fmt.Errorf("failed to write signature file: %w", err)
|
||||
}
|
||||
return nil
|
||||
}
|
||||
@@ -1,74 +0,0 @@
|
||||
package main
|
||||
|
||||
import (
|
||||
"fmt"
|
||||
"os"
|
||||
"time"
|
||||
|
||||
"github.com/spf13/cobra"
|
||||
|
||||
"github.com/netbirdio/netbird/client/internal/updatemanager/reposign"
|
||||
)
|
||||
|
||||
var (
|
||||
privKeyFile string
|
||||
pubKeyFile string
|
||||
rootExpiration time.Duration
|
||||
)
|
||||
|
||||
var createRootKeyCmd = &cobra.Command{
|
||||
Use: "create-root-key",
|
||||
Short: "Create a new root key pair",
|
||||
Long: `Create a new root key pair and specify an expiration time for it.`,
|
||||
SilenceUsage: true,
|
||||
RunE: func(cmd *cobra.Command, args []string) error {
|
||||
// Validate expiration
|
||||
if rootExpiration <= 0 {
|
||||
return fmt.Errorf("--expiration must be a positive duration (e.g., 720h, 365d, 8760h)")
|
||||
}
|
||||
|
||||
// Run main logic
|
||||
if err := handleGenerateRootKey(cmd, privKeyFile, pubKeyFile, rootExpiration); err != nil {
|
||||
return fmt.Errorf("failed to generate root key: %w", err)
|
||||
}
|
||||
return nil
|
||||
},
|
||||
}
|
||||
|
||||
func init() {
|
||||
rootCmd.AddCommand(createRootKeyCmd)
|
||||
createRootKeyCmd.Flags().StringVar(&privKeyFile, "priv-key-file", "", "Path to output private key file")
|
||||
createRootKeyCmd.Flags().StringVar(&pubKeyFile, "pub-key-file", "", "Path to output public key file")
|
||||
createRootKeyCmd.Flags().DurationVar(&rootExpiration, "expiration", 0, "Expiration time for the root key (e.g., 720h,)")
|
||||
|
||||
if err := createRootKeyCmd.MarkFlagRequired("priv-key-file"); err != nil {
|
||||
panic(err)
|
||||
}
|
||||
if err := createRootKeyCmd.MarkFlagRequired("pub-key-file"); err != nil {
|
||||
panic(err)
|
||||
}
|
||||
if err := createRootKeyCmd.MarkFlagRequired("expiration"); err != nil {
|
||||
panic(err)
|
||||
}
|
||||
}
|
||||
|
||||
func handleGenerateRootKey(cmd *cobra.Command, privKeyFile, pubKeyFile string, expiration time.Duration) error {
|
||||
rk, privPEM, pubPEM, err := reposign.GenerateRootKey(expiration)
|
||||
if err != nil {
|
||||
return fmt.Errorf("generate root key: %w", err)
|
||||
}
|
||||
|
||||
// Write private key
|
||||
if err := os.WriteFile(privKeyFile, privPEM, 0o600); err != nil {
|
||||
return fmt.Errorf("write private key file (%s): %w", privKeyFile, err)
|
||||
}
|
||||
|
||||
// Write public key
|
||||
if err := os.WriteFile(pubKeyFile, pubPEM, 0o600); err != nil {
|
||||
return fmt.Errorf("write public key file (%s): %w", pubKeyFile, err)
|
||||
}
|
||||
|
||||
cmd.Printf("%s\n\n", rk.String())
|
||||
cmd.Printf("✅ Root key pair generated successfully.\n")
|
||||
return nil
|
||||
}
|
||||
@@ -634,11 +634,7 @@ func parseAndStartLocalForward(ctx context.Context, c *sshclient.Client, forward
|
||||
return err
|
||||
}
|
||||
|
||||
if err := validateDestinationPort(remoteAddr); err != nil {
|
||||
return fmt.Errorf("invalid remote address: %w", err)
|
||||
}
|
||||
|
||||
log.Debugf("Local port forwarding: %s -> %s", localAddr, remoteAddr)
|
||||
cmd.Printf("Local port forwarding: %s -> %s\n", localAddr, remoteAddr)
|
||||
|
||||
go func() {
|
||||
if err := c.LocalPortForward(ctx, localAddr, remoteAddr); err != nil && !errors.Is(err, context.Canceled) {
|
||||
@@ -656,11 +652,7 @@ func parseAndStartRemoteForward(ctx context.Context, c *sshclient.Client, forwar
|
||||
return err
|
||||
}
|
||||
|
||||
if err := validateDestinationPort(localAddr); err != nil {
|
||||
return fmt.Errorf("invalid local address: %w", err)
|
||||
}
|
||||
|
||||
log.Debugf("Remote port forwarding: %s -> %s", remoteAddr, localAddr)
|
||||
cmd.Printf("Remote port forwarding: %s -> %s\n", remoteAddr, localAddr)
|
||||
|
||||
go func() {
|
||||
if err := c.RemotePortForward(ctx, remoteAddr, localAddr); err != nil && !errors.Is(err, context.Canceled) {
|
||||
@@ -671,35 +663,6 @@ func parseAndStartRemoteForward(ctx context.Context, c *sshclient.Client, forwar
|
||||
return nil
|
||||
}
|
||||
|
||||
// validateDestinationPort checks that the destination address has a valid port.
|
||||
// Port 0 is only valid for bind addresses (where the OS picks an available port),
|
||||
// not for destination addresses where we need to connect.
|
||||
func validateDestinationPort(addr string) error {
|
||||
if strings.HasPrefix(addr, "/") || strings.HasPrefix(addr, "./") {
|
||||
return nil
|
||||
}
|
||||
|
||||
_, portStr, err := net.SplitHostPort(addr)
|
||||
if err != nil {
|
||||
return fmt.Errorf("parse address %s: %w", addr, err)
|
||||
}
|
||||
|
||||
port, err := strconv.Atoi(portStr)
|
||||
if err != nil {
|
||||
return fmt.Errorf("invalid port %s: %w", portStr, err)
|
||||
}
|
||||
|
||||
if port == 0 {
|
||||
return fmt.Errorf("port 0 is not valid for destination address")
|
||||
}
|
||||
|
||||
if port < 0 || port > 65535 {
|
||||
return fmt.Errorf("port %d out of range (1-65535)", port)
|
||||
}
|
||||
|
||||
return nil
|
||||
}
|
||||
|
||||
// parsePortForwardSpec parses port forward specifications like "8080:localhost:80" or "[::1]:8080:localhost:80".
|
||||
// Also supports Unix sockets like "8080:/tmp/socket" or "127.0.0.1:8080:/tmp/socket".
|
||||
func parsePortForwardSpec(spec string) (string, string, error) {
|
||||
|
||||
@@ -103,13 +103,13 @@ func statusFunc(cmd *cobra.Command, args []string) error {
|
||||
var statusOutputString string
|
||||
switch {
|
||||
case detailFlag:
|
||||
statusOutputString = outputInformationHolder.FullDetailSummary()
|
||||
statusOutputString = nbstatus.ParseToFullDetailSummary(outputInformationHolder)
|
||||
case jsonFlag:
|
||||
statusOutputString, err = outputInformationHolder.JSON()
|
||||
statusOutputString, err = nbstatus.ParseToJSON(outputInformationHolder)
|
||||
case yamlFlag:
|
||||
statusOutputString, err = outputInformationHolder.YAML()
|
||||
statusOutputString, err = nbstatus.ParseToYAML(outputInformationHolder)
|
||||
default:
|
||||
statusOutputString = outputInformationHolder.GeneralSummary(false, false, false, false)
|
||||
statusOutputString = nbstatus.ParseGeneralSummary(outputInformationHolder, false, false, false, false)
|
||||
}
|
||||
|
||||
if err != nil {
|
||||
@@ -124,7 +124,6 @@ func statusFunc(cmd *cobra.Command, args []string) error {
|
||||
func getStatus(ctx context.Context, shouldRunProbes bool) (*proto.StatusResponse, error) {
|
||||
conn, err := DialClientGRPCServer(ctx, daemonAddr)
|
||||
if err != nil {
|
||||
//nolint
|
||||
return nil, fmt.Errorf("failed to connect to daemon error: %v\n"+
|
||||
"If the daemon is not running please run: "+
|
||||
"\nnetbird service install \nnetbird service start\n", err)
|
||||
|
||||
@@ -89,6 +89,9 @@ func startManagement(t *testing.T, config *config.Config, testFile string) (*grp
|
||||
t.Cleanup(cleanUp)
|
||||
|
||||
eventStore := &activity.InMemoryEventStore{}
|
||||
if err != nil {
|
||||
return nil, nil
|
||||
}
|
||||
|
||||
ctrl := gomock.NewController(t)
|
||||
t.Cleanup(ctrl.Finish)
|
||||
@@ -124,7 +127,7 @@ func startManagement(t *testing.T, config *config.Config, testFile string) (*grp
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
mgmtServer, err := nbgrpc.NewServer(config, accountManager, settingsMockManager, secretsManager, nil, nil, &mgmt.MockIntegratedValidator{}, networkMapController, nil)
|
||||
mgmtServer, err := nbgrpc.NewServer(config, accountManager, settingsMockManager, secretsManager, nil, nil, &mgmt.MockIntegratedValidator{}, networkMapController)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
|
||||
@@ -197,7 +197,7 @@ func runInForegroundMode(ctx context.Context, cmd *cobra.Command, activeProf *pr
|
||||
r := peer.NewRecorder(config.ManagementURL.String())
|
||||
r.GetFullStatus()
|
||||
|
||||
connectClient := internal.NewConnectClient(ctx, config, r, false)
|
||||
connectClient := internal.NewConnectClient(ctx, config, r)
|
||||
SetupDebugHandler(ctx, config, r, connectClient, "")
|
||||
|
||||
return connectClient.Run(nil)
|
||||
@@ -216,7 +216,6 @@ func runInDaemonMode(ctx context.Context, cmd *cobra.Command, pm *profilemanager
|
||||
|
||||
conn, err := DialClientGRPCServer(ctx, daemonAddr)
|
||||
if err != nil {
|
||||
//nolint
|
||||
return fmt.Errorf("failed to connect to daemon error: %v\n"+
|
||||
"If the daemon is not running please run: "+
|
||||
"\nnetbird service install \nnetbird service start\n", err)
|
||||
|
||||
@@ -1,13 +0,0 @@
|
||||
//go:build !windows && !darwin
|
||||
|
||||
package cmd
|
||||
|
||||
import (
|
||||
"github.com/spf13/cobra"
|
||||
)
|
||||
|
||||
var updateCmd *cobra.Command
|
||||
|
||||
func isUpdateBinary() bool {
|
||||
return false
|
||||
}
|
||||
@@ -1,75 +0,0 @@
|
||||
//go:build windows || darwin
|
||||
|
||||
package cmd
|
||||
|
||||
import (
|
||||
"context"
|
||||
"os"
|
||||
"path/filepath"
|
||||
"strings"
|
||||
|
||||
log "github.com/sirupsen/logrus"
|
||||
"github.com/spf13/cobra"
|
||||
|
||||
"github.com/netbirdio/netbird/client/internal/updatemanager/installer"
|
||||
"github.com/netbirdio/netbird/util"
|
||||
)
|
||||
|
||||
var (
|
||||
updateCmd = &cobra.Command{
|
||||
Use: "update",
|
||||
Short: "Update the NetBird client application",
|
||||
RunE: updateFunc,
|
||||
}
|
||||
|
||||
tempDirFlag string
|
||||
installerFile string
|
||||
serviceDirFlag string
|
||||
dryRunFlag bool
|
||||
)
|
||||
|
||||
func init() {
|
||||
updateCmd.Flags().StringVar(&tempDirFlag, "temp-dir", "", "temporary dir")
|
||||
updateCmd.Flags().StringVar(&installerFile, "installer-file", "", "installer file")
|
||||
updateCmd.Flags().StringVar(&serviceDirFlag, "service-dir", "", "service directory")
|
||||
updateCmd.Flags().BoolVar(&dryRunFlag, "dry-run", false, "dry run the update process without making any changes")
|
||||
}
|
||||
|
||||
// isUpdateBinary checks if the current executable is named "update" or "update.exe"
|
||||
func isUpdateBinary() bool {
|
||||
// Remove extension for cross-platform compatibility
|
||||
execPath, err := os.Executable()
|
||||
if err != nil {
|
||||
return false
|
||||
}
|
||||
baseName := filepath.Base(execPath)
|
||||
name := strings.TrimSuffix(baseName, filepath.Ext(baseName))
|
||||
|
||||
return name == installer.UpdaterBinaryNameWithoutExtension()
|
||||
}
|
||||
|
||||
func updateFunc(cmd *cobra.Command, args []string) error {
|
||||
if err := setupLogToFile(tempDirFlag); err != nil {
|
||||
return err
|
||||
}
|
||||
|
||||
log.Infof("updater started: %s", serviceDirFlag)
|
||||
updater := installer.NewWithDir(tempDirFlag)
|
||||
if err := updater.Setup(context.Background(), dryRunFlag, installerFile, serviceDirFlag); err != nil {
|
||||
log.Errorf("failed to update application: %v", err)
|
||||
return err
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
func setupLogToFile(dir string) error {
|
||||
logFile := filepath.Join(dir, installer.LogFile)
|
||||
|
||||
if _, err := os.Stat(logFile); err == nil {
|
||||
if err := os.Remove(logFile); err != nil {
|
||||
log.Errorf("failed to remove existing log file: %v\n", err)
|
||||
}
|
||||
}
|
||||
|
||||
return util.InitLog(logLevel, util.LogConsole, logFile)
|
||||
}
|
||||
@@ -20,7 +20,6 @@ import (
|
||||
"github.com/netbirdio/netbird/client/internal/profilemanager"
|
||||
sshcommon "github.com/netbirdio/netbird/client/ssh"
|
||||
"github.com/netbirdio/netbird/client/system"
|
||||
mgmProto "github.com/netbirdio/netbird/shared/management/proto"
|
||||
)
|
||||
|
||||
var (
|
||||
@@ -39,7 +38,6 @@ type Client struct {
|
||||
setupKey string
|
||||
jwtToken string
|
||||
connect *internal.ConnectClient
|
||||
recorder *peer.Status
|
||||
}
|
||||
|
||||
// Options configures a new Client.
|
||||
@@ -163,17 +161,11 @@ func New(opts Options) (*Client, error) {
|
||||
func (c *Client) Start(startCtx context.Context) error {
|
||||
c.mu.Lock()
|
||||
defer c.mu.Unlock()
|
||||
if c.connect != nil {
|
||||
if c.cancel != nil {
|
||||
return ErrClientAlreadyStarted
|
||||
}
|
||||
|
||||
ctx, cancel := context.WithCancel(internal.CtxInitState(context.Background()))
|
||||
defer func() {
|
||||
if c.connect == nil {
|
||||
cancel()
|
||||
}
|
||||
}()
|
||||
|
||||
ctx := internal.CtxInitState(context.Background())
|
||||
// nolint:staticcheck
|
||||
ctx = context.WithValue(ctx, system.DeviceNameCtxKey, c.deviceName)
|
||||
if err := internal.Login(ctx, c.config, c.setupKey, c.jwtToken); err != nil {
|
||||
@@ -181,9 +173,7 @@ func (c *Client) Start(startCtx context.Context) error {
|
||||
}
|
||||
|
||||
recorder := peer.NewRecorder(c.config.ManagementURL.String())
|
||||
c.recorder = recorder
|
||||
client := internal.NewConnectClient(ctx, c.config, recorder, false)
|
||||
client.SetSyncResponsePersistence(true)
|
||||
client := internal.NewConnectClient(ctx, c.config, recorder)
|
||||
|
||||
// either startup error (permanent backoff err) or nil err (successful engine up)
|
||||
// TODO: make after-startup backoff err available
|
||||
@@ -207,7 +197,6 @@ func (c *Client) Start(startCtx context.Context) error {
|
||||
}
|
||||
|
||||
c.connect = client
|
||||
c.cancel = cancel
|
||||
|
||||
return nil
|
||||
}
|
||||
@@ -222,23 +211,17 @@ func (c *Client) Stop(ctx context.Context) error {
|
||||
return ErrClientNotStarted
|
||||
}
|
||||
|
||||
if c.cancel != nil {
|
||||
c.cancel()
|
||||
c.cancel = nil
|
||||
}
|
||||
|
||||
done := make(chan error, 1)
|
||||
connect := c.connect
|
||||
go func() {
|
||||
done <- connect.Stop()
|
||||
done <- c.connect.Stop()
|
||||
}()
|
||||
|
||||
select {
|
||||
case <-ctx.Done():
|
||||
c.connect = nil
|
||||
c.cancel = nil
|
||||
return ctx.Err()
|
||||
case err := <-done:
|
||||
c.connect = nil
|
||||
c.cancel = nil
|
||||
if err != nil {
|
||||
return fmt.Errorf("stop: %w", err)
|
||||
}
|
||||
@@ -332,62 +315,6 @@ func (c *Client) NewHTTPClient() *http.Client {
|
||||
}
|
||||
}
|
||||
|
||||
// Status returns the current status of the client.
|
||||
func (c *Client) Status() (peer.FullStatus, error) {
|
||||
c.mu.Lock()
|
||||
recorder := c.recorder
|
||||
connect := c.connect
|
||||
c.mu.Unlock()
|
||||
|
||||
if recorder == nil {
|
||||
return peer.FullStatus{}, errors.New("client not started")
|
||||
}
|
||||
|
||||
if connect != nil {
|
||||
engine := connect.Engine()
|
||||
if engine != nil {
|
||||
_ = engine.RunHealthProbes(false)
|
||||
}
|
||||
}
|
||||
|
||||
return recorder.GetFullStatus(), nil
|
||||
}
|
||||
|
||||
// GetLatestSyncResponse returns the latest sync response from the management server.
|
||||
func (c *Client) GetLatestSyncResponse() (*mgmProto.SyncResponse, error) {
|
||||
engine, err := c.getEngine()
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
|
||||
syncResp, err := engine.GetLatestSyncResponse()
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("get sync response: %w", err)
|
||||
}
|
||||
|
||||
return syncResp, nil
|
||||
}
|
||||
|
||||
// SetLogLevel sets the logging level for the client and its components.
|
||||
func (c *Client) SetLogLevel(levelStr string) error {
|
||||
level, err := logrus.ParseLevel(levelStr)
|
||||
if err != nil {
|
||||
return fmt.Errorf("parse log level: %w", err)
|
||||
}
|
||||
|
||||
logrus.SetLevel(level)
|
||||
|
||||
c.mu.Lock()
|
||||
connect := c.connect
|
||||
c.mu.Unlock()
|
||||
|
||||
if connect != nil {
|
||||
connect.SetLogLevel(level)
|
||||
}
|
||||
|
||||
return nil
|
||||
}
|
||||
|
||||
// VerifySSHHostKey verifies an SSH host key against stored peer keys.
|
||||
// Returns nil if the key matches, ErrPeerNotFound if peer is not in network,
|
||||
// ErrNoStoredKey if peer has no stored key, or an error for verification failures.
|
||||
|
||||
@@ -386,8 +386,11 @@ func (m *aclManager) updateState() {
|
||||
|
||||
// filterRuleSpecs returns the specs of a filtering rule
|
||||
func filterRuleSpecs(ip net.IP, protocol string, sPort, dPort *firewall.Port, action firewall.Action, ipsetName string) (specs []string) {
|
||||
matchByIP := true
|
||||
// don't use IP matching if IP is 0.0.0.0
|
||||
matchByIP := !ip.IsUnspecified()
|
||||
if ip.IsUnspecified() {
|
||||
matchByIP = false
|
||||
}
|
||||
|
||||
if matchByIP {
|
||||
if ipsetName != "" {
|
||||
|
||||
@@ -161,7 +161,7 @@ func TestIptablesManagerDenyRules(t *testing.T) {
|
||||
t.Logf(" [%d] %s", i, rule)
|
||||
}
|
||||
|
||||
var denyRuleIndex, acceptRuleIndex = -1, -1
|
||||
var denyRuleIndex, acceptRuleIndex int = -1, -1
|
||||
for i, rule := range rules {
|
||||
if strings.Contains(rule, "DROP") {
|
||||
t.Logf("Found DROP rule at index %d: %s", i, rule)
|
||||
|
||||
@@ -198,7 +198,7 @@ func TestNftablesManagerRuleOrder(t *testing.T) {
|
||||
t.Logf("Found %d rules in nftables chain", len(rules))
|
||||
|
||||
// Find the accept and deny rules and verify deny comes before accept
|
||||
var acceptRuleIndex, denyRuleIndex = -1, -1
|
||||
var acceptRuleIndex, denyRuleIndex int = -1, -1
|
||||
for i, rule := range rules {
|
||||
hasAcceptHTTPSet := false
|
||||
hasDenyHTTPSet := false
|
||||
@@ -208,13 +208,11 @@ func TestNftablesManagerRuleOrder(t *testing.T) {
|
||||
for _, e := range rule.Exprs {
|
||||
// Check for set lookup
|
||||
if lookup, ok := e.(*expr.Lookup); ok {
|
||||
switch lookup.SetName {
|
||||
case "accept-http":
|
||||
if lookup.SetName == "accept-http" {
|
||||
hasAcceptHTTPSet = true
|
||||
case "deny-http":
|
||||
} else if lookup.SetName == "deny-http" {
|
||||
hasDenyHTTPSet = true
|
||||
}
|
||||
|
||||
}
|
||||
// Check for port 80
|
||||
if cmp, ok := e.(*expr.Cmp); ok {
|
||||
@@ -224,10 +222,9 @@ func TestNftablesManagerRuleOrder(t *testing.T) {
|
||||
}
|
||||
// Check for verdict
|
||||
if verdict, ok := e.(*expr.Verdict); ok {
|
||||
switch verdict.Kind {
|
||||
case expr.VerdictAccept:
|
||||
if verdict.Kind == expr.VerdictAccept {
|
||||
action = "ACCEPT"
|
||||
case expr.VerdictDrop:
|
||||
} else if verdict.Kind == expr.VerdictDrop {
|
||||
action = "DROP"
|
||||
}
|
||||
}
|
||||
@@ -389,97 +386,6 @@ func TestNftablesManagerCompatibilityWithIptables(t *testing.T) {
|
||||
verifyIptablesOutput(t, stdout, stderr)
|
||||
}
|
||||
|
||||
func TestNftablesManagerCompatibilityWithIptablesFor6kPrefixes(t *testing.T) {
|
||||
if check() != NFTABLES {
|
||||
t.Skip("nftables not supported on this system")
|
||||
}
|
||||
|
||||
if _, err := exec.LookPath("iptables-save"); err != nil {
|
||||
t.Skipf("iptables-save not available on this system: %v", err)
|
||||
}
|
||||
|
||||
// First ensure iptables-nft tables exist by running iptables-save
|
||||
stdout, stderr := runIptablesSave(t)
|
||||
verifyIptablesOutput(t, stdout, stderr)
|
||||
|
||||
manager, err := Create(ifaceMock, iface.DefaultMTU)
|
||||
require.NoError(t, err, "failed to create manager")
|
||||
require.NoError(t, manager.Init(nil))
|
||||
|
||||
t.Cleanup(func() {
|
||||
err := manager.Close(nil)
|
||||
require.NoError(t, err, "failed to reset manager state")
|
||||
|
||||
// Verify iptables output after reset
|
||||
stdout, stderr := runIptablesSave(t)
|
||||
verifyIptablesOutput(t, stdout, stderr)
|
||||
})
|
||||
|
||||
const octet2Count = 25
|
||||
const octet3Count = 255
|
||||
prefixes := make([]netip.Prefix, 0, (octet2Count-1)*(octet3Count-1))
|
||||
for i := 1; i < octet2Count; i++ {
|
||||
for j := 1; j < octet3Count; j++ {
|
||||
addr := netip.AddrFrom4([4]byte{192, byte(j), byte(i), 0})
|
||||
prefixes = append(prefixes, netip.PrefixFrom(addr, 24))
|
||||
}
|
||||
}
|
||||
_, err = manager.AddRouteFiltering(
|
||||
nil,
|
||||
prefixes,
|
||||
fw.Network{Prefix: netip.MustParsePrefix("10.2.0.0/24")},
|
||||
fw.ProtocolTCP,
|
||||
nil,
|
||||
&fw.Port{Values: []uint16{443}},
|
||||
fw.ActionAccept,
|
||||
)
|
||||
require.NoError(t, err, "failed to add route filtering rule")
|
||||
|
||||
stdout, stderr = runIptablesSave(t)
|
||||
verifyIptablesOutput(t, stdout, stderr)
|
||||
}
|
||||
|
||||
func TestNftablesManagerCompatibilityWithIptablesForEmptyPrefixes(t *testing.T) {
|
||||
if check() != NFTABLES {
|
||||
t.Skip("nftables not supported on this system")
|
||||
}
|
||||
|
||||
if _, err := exec.LookPath("iptables-save"); err != nil {
|
||||
t.Skipf("iptables-save not available on this system: %v", err)
|
||||
}
|
||||
|
||||
// First ensure iptables-nft tables exist by running iptables-save
|
||||
stdout, stderr := runIptablesSave(t)
|
||||
verifyIptablesOutput(t, stdout, stderr)
|
||||
|
||||
manager, err := Create(ifaceMock, iface.DefaultMTU)
|
||||
require.NoError(t, err, "failed to create manager")
|
||||
require.NoError(t, manager.Init(nil))
|
||||
|
||||
t.Cleanup(func() {
|
||||
err := manager.Close(nil)
|
||||
require.NoError(t, err, "failed to reset manager state")
|
||||
|
||||
// Verify iptables output after reset
|
||||
stdout, stderr := runIptablesSave(t)
|
||||
verifyIptablesOutput(t, stdout, stderr)
|
||||
})
|
||||
|
||||
_, err = manager.AddRouteFiltering(
|
||||
nil,
|
||||
[]netip.Prefix{},
|
||||
fw.Network{Prefix: netip.MustParsePrefix("10.2.0.0/24")},
|
||||
fw.ProtocolTCP,
|
||||
nil,
|
||||
&fw.Port{Values: []uint16{443}},
|
||||
fw.ActionAccept,
|
||||
)
|
||||
require.NoError(t, err, "failed to add route filtering rule")
|
||||
|
||||
stdout, stderr = runIptablesSave(t)
|
||||
verifyIptablesOutput(t, stdout, stderr)
|
||||
}
|
||||
|
||||
func compareExprsIgnoringCounters(t *testing.T, got, want []expr.Any) {
|
||||
t.Helper()
|
||||
require.Equal(t, len(got), len(want), "expression count mismatch")
|
||||
|
||||
@@ -48,12 +48,10 @@ const (
|
||||
|
||||
// ipTCPHeaderMinSize represents minimum IP (20) + TCP (20) header size for MSS calculation
|
||||
ipTCPHeaderMinSize = 40
|
||||
|
||||
// maxPrefixesSet 1638 prefixes start to fail, taking some margin
|
||||
maxPrefixesSet = 1500
|
||||
refreshRulesMapError = "refresh rules map: %w"
|
||||
)
|
||||
|
||||
const refreshRulesMapError = "refresh rules map: %w"
|
||||
|
||||
var (
|
||||
errFilterTableNotFound = fmt.Errorf("'filter' table not found")
|
||||
)
|
||||
@@ -515,35 +513,16 @@ func (r *router) createIpSet(setName string, input setInput) (*nftables.Set, err
|
||||
}
|
||||
|
||||
elements := convertPrefixesToSet(prefixes)
|
||||
nElements := len(elements)
|
||||
|
||||
maxElements := maxPrefixesSet * 2
|
||||
initialElements := elements[:min(maxElements, nElements)]
|
||||
|
||||
if err := r.conn.AddSet(nfset, initialElements); err != nil {
|
||||
return nil, fmt.Errorf("error adding set %s: %w", setName, err)
|
||||
if err := r.conn.AddSet(nfset, elements); err != nil {
|
||||
return nil, fmt.Errorf("error adding elements to set %s: %w", setName, err)
|
||||
}
|
||||
|
||||
if err := r.conn.Flush(); err != nil {
|
||||
return nil, fmt.Errorf("flush error: %w", err)
|
||||
}
|
||||
log.Debugf("Created new ipset: %s with %d initial prefixes (total prefixes %d)", setName, len(initialElements)/2, len(prefixes))
|
||||
|
||||
var subEnd int
|
||||
for subStart := maxElements; subStart < nElements; subStart += maxElements {
|
||||
subEnd = min(subStart+maxElements, nElements)
|
||||
subElement := elements[subStart:subEnd]
|
||||
nSubPrefixes := len(subElement) / 2
|
||||
log.Tracef("Adding new prefixes (%d) in ipset: %s", nSubPrefixes, setName)
|
||||
if err := r.conn.SetAddElements(nfset, subElement); err != nil {
|
||||
return nil, fmt.Errorf("error adding prefixes (%d) to set %s: %w", nSubPrefixes, setName, err)
|
||||
}
|
||||
if err := r.conn.Flush(); err != nil {
|
||||
return nil, fmt.Errorf("flush error: %w", err)
|
||||
}
|
||||
log.Debugf("Added new prefixes (%d) in ipset: %s", nSubPrefixes, setName)
|
||||
}
|
||||
log.Printf("Created new ipset: %s with %d elements", setName, len(elements)/2)
|
||||
|
||||
log.Infof("Created new ipset: %s with %d prefixes", setName, len(prefixes))
|
||||
return nfset, nil
|
||||
}
|
||||
|
||||
|
||||
@@ -29,7 +29,7 @@ import (
|
||||
)
|
||||
|
||||
const (
|
||||
layerTypeAll = 255
|
||||
layerTypeAll = 0
|
||||
|
||||
// ipTCPHeaderMinSize represents minimum IP (20) + TCP (20) header size for MSS calculation
|
||||
ipTCPHeaderMinSize = 40
|
||||
@@ -262,7 +262,10 @@ func create(iface common.IFaceMapper, nativeFirewall firewall.Manager, disableSe
|
||||
}
|
||||
|
||||
func (m *Manager) blockInvalidRouted(iface common.IFaceMapper) (firewall.Rule, error) {
|
||||
wgPrefix := iface.Address().Network
|
||||
wgPrefix, err := netip.ParsePrefix(iface.Address().Network.String())
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("parse wireguard network: %w", err)
|
||||
}
|
||||
log.Debugf("blocking invalid routed traffic for %s", wgPrefix)
|
||||
|
||||
rule, err := m.addRouteFiltering(
|
||||
@@ -436,7 +439,19 @@ func (m *Manager) AddPeerFiltering(
|
||||
r.sPort = sPort
|
||||
r.dPort = dPort
|
||||
|
||||
r.protoLayer = protoToLayer(proto, r.ipLayer)
|
||||
switch proto {
|
||||
case firewall.ProtocolTCP:
|
||||
r.protoLayer = layers.LayerTypeTCP
|
||||
case firewall.ProtocolUDP:
|
||||
r.protoLayer = layers.LayerTypeUDP
|
||||
case firewall.ProtocolICMP:
|
||||
r.protoLayer = layers.LayerTypeICMPv4
|
||||
if r.ipLayer == layers.LayerTypeIPv6 {
|
||||
r.protoLayer = layers.LayerTypeICMPv6
|
||||
}
|
||||
case firewall.ProtocolALL:
|
||||
r.protoLayer = layerTypeAll
|
||||
}
|
||||
|
||||
m.mutex.Lock()
|
||||
var targetMap map[netip.Addr]RuleSet
|
||||
@@ -481,17 +496,16 @@ func (m *Manager) addRouteFiltering(
|
||||
}
|
||||
|
||||
ruleID := uuid.New().String()
|
||||
|
||||
rule := RouteRule{
|
||||
// TODO: consolidate these IDs
|
||||
id: ruleID,
|
||||
mgmtId: id,
|
||||
sources: sources,
|
||||
dstSet: destination.Set,
|
||||
protoLayer: protoToLayer(proto, layers.LayerTypeIPv4),
|
||||
srcPort: sPort,
|
||||
dstPort: dPort,
|
||||
action: action,
|
||||
id: ruleID,
|
||||
mgmtId: id,
|
||||
sources: sources,
|
||||
dstSet: destination.Set,
|
||||
proto: proto,
|
||||
srcPort: sPort,
|
||||
dstPort: dPort,
|
||||
action: action,
|
||||
}
|
||||
if destination.IsPrefix() {
|
||||
rule.destinations = []netip.Prefix{destination.Prefix}
|
||||
@@ -781,7 +795,7 @@ func (m *Manager) recalculateTCPChecksum(packetData []byte, d *decoder, tcpHeade
|
||||
pseudoSum += uint32(d.ip4.Protocol)
|
||||
pseudoSum += uint32(tcpLength)
|
||||
|
||||
var sum = pseudoSum
|
||||
var sum uint32 = pseudoSum
|
||||
for i := 0; i < tcpLength-1; i += 2 {
|
||||
sum += uint32(tcpLayer[i])<<8 | uint32(tcpLayer[i+1])
|
||||
}
|
||||
@@ -931,7 +945,7 @@ func (m *Manager) filterInbound(packetData []byte, size int) bool {
|
||||
func (m *Manager) handleLocalTraffic(d *decoder, srcIP, dstIP netip.Addr, packetData []byte, size int) bool {
|
||||
ruleID, blocked := m.peerACLsBlock(srcIP, d, packetData)
|
||||
if blocked {
|
||||
pnum := getProtocolFromPacket(d)
|
||||
_, pnum := getProtocolFromPacket(d)
|
||||
srcPort, dstPort := getPortsFromPacket(d)
|
||||
|
||||
m.logger.Trace6("Dropping local packet (ACL denied): rule_id=%s proto=%v src=%s:%d dst=%s:%d",
|
||||
@@ -996,22 +1010,20 @@ func (m *Manager) handleRoutedTraffic(d *decoder, srcIP, dstIP netip.Addr, packe
|
||||
return false
|
||||
}
|
||||
|
||||
protoLayer := d.decoded[1]
|
||||
proto, pnum := getProtocolFromPacket(d)
|
||||
srcPort, dstPort := getPortsFromPacket(d)
|
||||
|
||||
ruleID, pass := m.routeACLsPass(srcIP, dstIP, protoLayer, srcPort, dstPort)
|
||||
ruleID, pass := m.routeACLsPass(srcIP, dstIP, proto, srcPort, dstPort)
|
||||
if !pass {
|
||||
proto := getProtocolFromPacket(d)
|
||||
|
||||
m.logger.Trace6("Dropping routed packet (ACL denied): rule_id=%s proto=%v src=%s:%d dst=%s:%d",
|
||||
ruleID, proto, srcIP, srcPort, dstIP, dstPort)
|
||||
ruleID, pnum, srcIP, srcPort, dstIP, dstPort)
|
||||
|
||||
m.flowLogger.StoreEvent(nftypes.EventFields{
|
||||
FlowID: uuid.New(),
|
||||
Type: nftypes.TypeDrop,
|
||||
RuleID: ruleID,
|
||||
Direction: nftypes.Ingress,
|
||||
Protocol: proto,
|
||||
Protocol: pnum,
|
||||
SourceIP: srcIP,
|
||||
DestIP: dstIP,
|
||||
SourcePort: srcPort,
|
||||
@@ -1040,33 +1052,16 @@ func (m *Manager) handleRoutedTraffic(d *decoder, srcIP, dstIP netip.Addr, packe
|
||||
return true
|
||||
}
|
||||
|
||||
func protoToLayer(proto firewall.Protocol, ipLayer gopacket.LayerType) gopacket.LayerType {
|
||||
switch proto {
|
||||
case firewall.ProtocolTCP:
|
||||
return layers.LayerTypeTCP
|
||||
case firewall.ProtocolUDP:
|
||||
return layers.LayerTypeUDP
|
||||
case firewall.ProtocolICMP:
|
||||
if ipLayer == layers.LayerTypeIPv6 {
|
||||
return layers.LayerTypeICMPv6
|
||||
}
|
||||
return layers.LayerTypeICMPv4
|
||||
case firewall.ProtocolALL:
|
||||
return layerTypeAll
|
||||
}
|
||||
return 0
|
||||
}
|
||||
|
||||
func getProtocolFromPacket(d *decoder) nftypes.Protocol {
|
||||
func getProtocolFromPacket(d *decoder) (firewall.Protocol, nftypes.Protocol) {
|
||||
switch d.decoded[1] {
|
||||
case layers.LayerTypeTCP:
|
||||
return nftypes.TCP
|
||||
return firewall.ProtocolTCP, nftypes.TCP
|
||||
case layers.LayerTypeUDP:
|
||||
return nftypes.UDP
|
||||
return firewall.ProtocolUDP, nftypes.UDP
|
||||
case layers.LayerTypeICMPv4, layers.LayerTypeICMPv6:
|
||||
return nftypes.ICMP
|
||||
return firewall.ProtocolICMP, nftypes.ICMP
|
||||
default:
|
||||
return nftypes.ProtocolUnknown
|
||||
return firewall.ProtocolALL, nftypes.ProtocolUnknown
|
||||
}
|
||||
}
|
||||
|
||||
@@ -1238,30 +1233,19 @@ func validateRule(ip netip.Addr, packetData []byte, rules map[string]PeerRule, d
|
||||
}
|
||||
|
||||
// routeACLsPass returns true if the packet is allowed by the route ACLs
|
||||
func (m *Manager) routeACLsPass(srcIP, dstIP netip.Addr, protoLayer gopacket.LayerType, srcPort, dstPort uint16) ([]byte, bool) {
|
||||
func (m *Manager) routeACLsPass(srcIP, dstIP netip.Addr, proto firewall.Protocol, srcPort, dstPort uint16) ([]byte, bool) {
|
||||
m.mutex.RLock()
|
||||
defer m.mutex.RUnlock()
|
||||
|
||||
for _, rule := range m.routeRules {
|
||||
if matches := m.ruleMatches(rule, srcIP, dstIP, protoLayer, srcPort, dstPort); matches {
|
||||
if matches := m.ruleMatches(rule, srcIP, dstIP, proto, srcPort, dstPort); matches {
|
||||
return rule.mgmtId, rule.action == firewall.ActionAccept
|
||||
}
|
||||
}
|
||||
return nil, false
|
||||
}
|
||||
|
||||
func (m *Manager) ruleMatches(rule *RouteRule, srcAddr, dstAddr netip.Addr, protoLayer gopacket.LayerType, srcPort, dstPort uint16) bool {
|
||||
// TODO: handle ipv6 vs ipv4 icmp rules
|
||||
if rule.protoLayer != layerTypeAll && rule.protoLayer != protoLayer {
|
||||
return false
|
||||
}
|
||||
|
||||
if protoLayer == layers.LayerTypeTCP || protoLayer == layers.LayerTypeUDP {
|
||||
if !portsMatch(rule.srcPort, srcPort) || !portsMatch(rule.dstPort, dstPort) {
|
||||
return false
|
||||
}
|
||||
}
|
||||
|
||||
func (m *Manager) ruleMatches(rule *RouteRule, srcAddr, dstAddr netip.Addr, proto firewall.Protocol, srcPort, dstPort uint16) bool {
|
||||
destMatched := false
|
||||
for _, dst := range rule.destinations {
|
||||
if dst.Contains(dstAddr) {
|
||||
@@ -1280,8 +1264,21 @@ func (m *Manager) ruleMatches(rule *RouteRule, srcAddr, dstAddr netip.Addr, prot
|
||||
break
|
||||
}
|
||||
}
|
||||
if !sourceMatched {
|
||||
return false
|
||||
}
|
||||
|
||||
return sourceMatched
|
||||
if rule.proto != firewall.ProtocolALL && rule.proto != proto {
|
||||
return false
|
||||
}
|
||||
|
||||
if proto == firewall.ProtocolTCP || proto == firewall.ProtocolUDP {
|
||||
if !portsMatch(rule.srcPort, srcPort) || !portsMatch(rule.dstPort, dstPort) {
|
||||
return false
|
||||
}
|
||||
}
|
||||
|
||||
return true
|
||||
}
|
||||
|
||||
// AddUDPPacketHook calls hook when UDP packet from given direction matched
|
||||
|
||||
@@ -955,7 +955,7 @@ func BenchmarkRouteACLs(b *testing.B) {
|
||||
for _, tc := range cases {
|
||||
srcIP := netip.MustParseAddr(tc.srcIP)
|
||||
dstIP := netip.MustParseAddr(tc.dstIP)
|
||||
manager.routeACLsPass(srcIP, dstIP, protoToLayer(tc.proto, layers.LayerTypeIPv4), 0, tc.dstPort)
|
||||
manager.routeACLsPass(srcIP, dstIP, tc.proto, 0, tc.dstPort)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
@@ -1259,7 +1259,7 @@ func TestRouteACLFiltering(t *testing.T) {
|
||||
|
||||
// testing routeACLsPass only and not FilterInbound, as routed packets are dropped after being passed
|
||||
// to the forwarder
|
||||
_, isAllowed := manager.routeACLsPass(srcIP, dstIP, protoToLayer(tc.proto, layers.LayerTypeIPv4), tc.srcPort, tc.dstPort)
|
||||
_, isAllowed := manager.routeACLsPass(srcIP, dstIP, tc.proto, tc.srcPort, tc.dstPort)
|
||||
require.Equal(t, tc.shouldPass, isAllowed)
|
||||
})
|
||||
}
|
||||
@@ -1445,7 +1445,7 @@ func TestRouteACLOrder(t *testing.T) {
|
||||
srcIP := netip.MustParseAddr(p.srcIP)
|
||||
dstIP := netip.MustParseAddr(p.dstIP)
|
||||
|
||||
_, isAllowed := manager.routeACLsPass(srcIP, dstIP, protoToLayer(p.proto, layers.LayerTypeIPv4), p.srcPort, p.dstPort)
|
||||
_, isAllowed := manager.routeACLsPass(srcIP, dstIP, p.proto, p.srcPort, p.dstPort)
|
||||
require.Equal(t, p.shouldPass, isAllowed, "packet %d failed", i)
|
||||
}
|
||||
})
|
||||
@@ -1488,13 +1488,13 @@ func TestRouteACLSet(t *testing.T) {
|
||||
dstIP := netip.MustParseAddr("192.168.1.100")
|
||||
|
||||
// Check that traffic is dropped (empty set shouldn't match anything)
|
||||
_, isAllowed := manager.routeACLsPass(srcIP, dstIP, protoToLayer(fw.ProtocolTCP, layers.LayerTypeIPv4), 12345, 80)
|
||||
_, isAllowed := manager.routeACLsPass(srcIP, dstIP, fw.ProtocolTCP, 12345, 80)
|
||||
require.False(t, isAllowed, "Empty set should not allow any traffic")
|
||||
|
||||
err = manager.UpdateSet(set, []netip.Prefix{netip.MustParsePrefix("192.168.1.0/24")})
|
||||
require.NoError(t, err)
|
||||
|
||||
// Now the packet should be allowed
|
||||
_, isAllowed = manager.routeACLsPass(srcIP, dstIP, protoToLayer(fw.ProtocolTCP, layers.LayerTypeIPv4), 12345, 80)
|
||||
_, isAllowed = manager.routeACLsPass(srcIP, dstIP, fw.ProtocolTCP, 12345, 80)
|
||||
require.True(t, isAllowed, "After set update, traffic to the added network should be allowed")
|
||||
}
|
||||
|
||||
@@ -767,9 +767,9 @@ func TestUpdateSetMerge(t *testing.T) {
|
||||
dstIP2 := netip.MustParseAddr("192.168.1.100")
|
||||
dstIP3 := netip.MustParseAddr("172.16.0.100")
|
||||
|
||||
_, isAllowed1 := manager.routeACLsPass(srcIP, dstIP1, protoToLayer(fw.ProtocolTCP, layers.LayerTypeIPv4), 12345, 80)
|
||||
_, isAllowed2 := manager.routeACLsPass(srcIP, dstIP2, protoToLayer(fw.ProtocolTCP, layers.LayerTypeIPv4), 12345, 80)
|
||||
_, isAllowed3 := manager.routeACLsPass(srcIP, dstIP3, protoToLayer(fw.ProtocolTCP, layers.LayerTypeIPv4), 12345, 80)
|
||||
_, isAllowed1 := manager.routeACLsPass(srcIP, dstIP1, fw.ProtocolTCP, 12345, 80)
|
||||
_, isAllowed2 := manager.routeACLsPass(srcIP, dstIP2, fw.ProtocolTCP, 12345, 80)
|
||||
_, isAllowed3 := manager.routeACLsPass(srcIP, dstIP3, fw.ProtocolTCP, 12345, 80)
|
||||
|
||||
require.True(t, isAllowed1, "Traffic to 10.0.0.100 should be allowed")
|
||||
require.True(t, isAllowed2, "Traffic to 192.168.1.100 should be allowed")
|
||||
@@ -784,8 +784,8 @@ func TestUpdateSetMerge(t *testing.T) {
|
||||
require.NoError(t, err)
|
||||
|
||||
// Check that all original prefixes are still included
|
||||
_, isAllowed1 = manager.routeACLsPass(srcIP, dstIP1, protoToLayer(fw.ProtocolTCP, layers.LayerTypeIPv4), 12345, 80)
|
||||
_, isAllowed2 = manager.routeACLsPass(srcIP, dstIP2, protoToLayer(fw.ProtocolTCP, layers.LayerTypeIPv4), 12345, 80)
|
||||
_, isAllowed1 = manager.routeACLsPass(srcIP, dstIP1, fw.ProtocolTCP, 12345, 80)
|
||||
_, isAllowed2 = manager.routeACLsPass(srcIP, dstIP2, fw.ProtocolTCP, 12345, 80)
|
||||
require.True(t, isAllowed1, "Traffic to 10.0.0.100 should still be allowed after update")
|
||||
require.True(t, isAllowed2, "Traffic to 192.168.1.100 should still be allowed after update")
|
||||
|
||||
@@ -793,8 +793,8 @@ func TestUpdateSetMerge(t *testing.T) {
|
||||
dstIP4 := netip.MustParseAddr("172.16.1.100")
|
||||
dstIP5 := netip.MustParseAddr("10.1.0.50")
|
||||
|
||||
_, isAllowed4 := manager.routeACLsPass(srcIP, dstIP4, protoToLayer(fw.ProtocolTCP, layers.LayerTypeIPv4), 12345, 80)
|
||||
_, isAllowed5 := manager.routeACLsPass(srcIP, dstIP5, protoToLayer(fw.ProtocolTCP, layers.LayerTypeIPv4), 12345, 80)
|
||||
_, isAllowed4 := manager.routeACLsPass(srcIP, dstIP4, fw.ProtocolTCP, 12345, 80)
|
||||
_, isAllowed5 := manager.routeACLsPass(srcIP, dstIP5, fw.ProtocolTCP, 12345, 80)
|
||||
|
||||
require.True(t, isAllowed4, "Traffic to new prefix 172.16.0.0/16 should be allowed")
|
||||
require.True(t, isAllowed5, "Traffic to new prefix 10.1.0.0/24 should be allowed")
|
||||
@@ -922,7 +922,7 @@ func TestUpdateSetDeduplication(t *testing.T) {
|
||||
|
||||
srcIP := netip.MustParseAddr("100.10.0.1")
|
||||
for _, tc := range testCases {
|
||||
_, isAllowed := manager.routeACLsPass(srcIP, tc.dstIP, protoToLayer(fw.ProtocolTCP, layers.LayerTypeIPv4), 12345, 80)
|
||||
_, isAllowed := manager.routeACLsPass(srcIP, tc.dstIP, fw.ProtocolTCP, 12345, 80)
|
||||
require.Equal(t, tc.expected, isAllowed, tc.desc)
|
||||
}
|
||||
}
|
||||
|
||||
@@ -2,7 +2,6 @@ package forwarder
|
||||
|
||||
import (
|
||||
"fmt"
|
||||
"sync/atomic"
|
||||
|
||||
wgdevice "golang.zx2c4.com/wireguard/device"
|
||||
"gvisor.dev/gvisor/pkg/tcpip"
|
||||
@@ -17,7 +16,7 @@ type endpoint struct {
|
||||
logger *nblog.Logger
|
||||
dispatcher stack.NetworkDispatcher
|
||||
device *wgdevice.Device
|
||||
mtu atomic.Uint32
|
||||
mtu uint32
|
||||
}
|
||||
|
||||
func (e *endpoint) Attach(dispatcher stack.NetworkDispatcher) {
|
||||
@@ -29,7 +28,7 @@ func (e *endpoint) IsAttached() bool {
|
||||
}
|
||||
|
||||
func (e *endpoint) MTU() uint32 {
|
||||
return e.mtu.Load()
|
||||
return e.mtu
|
||||
}
|
||||
|
||||
func (e *endpoint) Capabilities() stack.LinkEndpointCapabilities {
|
||||
@@ -83,22 +82,6 @@ func (e *endpoint) ParseHeader(*stack.PacketBuffer) bool {
|
||||
return true
|
||||
}
|
||||
|
||||
func (e *endpoint) Close() {
|
||||
// Endpoint cleanup - nothing to do as device is managed externally
|
||||
}
|
||||
|
||||
func (e *endpoint) SetLinkAddress(tcpip.LinkAddress) {
|
||||
// Link address is not used for this endpoint type
|
||||
}
|
||||
|
||||
func (e *endpoint) SetMTU(mtu uint32) {
|
||||
e.mtu.Store(mtu)
|
||||
}
|
||||
|
||||
func (e *endpoint) SetOnCloseAction(func()) {
|
||||
// No action needed on close
|
||||
}
|
||||
|
||||
type epID stack.TransportEndpointID
|
||||
|
||||
func (i epID) String() string {
|
||||
|
||||
@@ -7,7 +7,6 @@ import (
|
||||
"net/netip"
|
||||
"runtime"
|
||||
"sync"
|
||||
"time"
|
||||
|
||||
log "github.com/sirupsen/logrus"
|
||||
"gvisor.dev/gvisor/pkg/buffer"
|
||||
@@ -36,16 +35,14 @@ type Forwarder struct {
|
||||
logger *nblog.Logger
|
||||
flowLogger nftypes.FlowLogger
|
||||
// ruleIdMap is used to store the rule ID for a given connection
|
||||
ruleIdMap sync.Map
|
||||
stack *stack.Stack
|
||||
endpoint *endpoint
|
||||
udpForwarder *udpForwarder
|
||||
ctx context.Context
|
||||
cancel context.CancelFunc
|
||||
ip tcpip.Address
|
||||
netstack bool
|
||||
hasRawICMPAccess bool
|
||||
pingSemaphore chan struct{}
|
||||
ruleIdMap sync.Map
|
||||
stack *stack.Stack
|
||||
endpoint *endpoint
|
||||
udpForwarder *udpForwarder
|
||||
ctx context.Context
|
||||
cancel context.CancelFunc
|
||||
ip tcpip.Address
|
||||
netstack bool
|
||||
}
|
||||
|
||||
func New(iface common.IFaceMapper, logger *nblog.Logger, flowLogger nftypes.FlowLogger, netstack bool, mtu uint16) (*Forwarder, error) {
|
||||
@@ -63,8 +60,8 @@ func New(iface common.IFaceMapper, logger *nblog.Logger, flowLogger nftypes.Flow
|
||||
endpoint := &endpoint{
|
||||
logger: logger,
|
||||
device: iface.GetWGDevice(),
|
||||
mtu: uint32(mtu),
|
||||
}
|
||||
endpoint.mtu.Store(uint32(mtu))
|
||||
|
||||
if err := s.CreateNIC(nicID, endpoint); err != nil {
|
||||
return nil, fmt.Errorf("create NIC: %v", err)
|
||||
@@ -106,16 +103,15 @@ func New(iface common.IFaceMapper, logger *nblog.Logger, flowLogger nftypes.Flow
|
||||
|
||||
ctx, cancel := context.WithCancel(context.Background())
|
||||
f := &Forwarder{
|
||||
logger: logger,
|
||||
flowLogger: flowLogger,
|
||||
stack: s,
|
||||
endpoint: endpoint,
|
||||
udpForwarder: newUDPForwarder(mtu, logger, flowLogger),
|
||||
ctx: ctx,
|
||||
cancel: cancel,
|
||||
netstack: netstack,
|
||||
ip: tcpip.AddrFromSlice(iface.Address().IP.AsSlice()),
|
||||
pingSemaphore: make(chan struct{}, 3),
|
||||
logger: logger,
|
||||
flowLogger: flowLogger,
|
||||
stack: s,
|
||||
endpoint: endpoint,
|
||||
udpForwarder: newUDPForwarder(mtu, logger, flowLogger),
|
||||
ctx: ctx,
|
||||
cancel: cancel,
|
||||
netstack: netstack,
|
||||
ip: tcpip.AddrFromSlice(iface.Address().IP.AsSlice()),
|
||||
}
|
||||
|
||||
receiveWindow := defaultReceiveWindow
|
||||
@@ -133,8 +129,6 @@ func New(iface common.IFaceMapper, logger *nblog.Logger, flowLogger nftypes.Flow
|
||||
|
||||
s.SetTransportProtocolHandler(icmp.ProtocolNumber4, f.handleICMP)
|
||||
|
||||
f.checkICMPCapability()
|
||||
|
||||
log.Debugf("forwarder: Initialization complete with NIC %d", nicID)
|
||||
return f, nil
|
||||
}
|
||||
@@ -204,24 +198,3 @@ func buildKey(srcIP, dstIP netip.Addr, srcPort, dstPort uint16) conntrack.ConnKe
|
||||
DstPort: dstPort,
|
||||
}
|
||||
}
|
||||
|
||||
// checkICMPCapability tests whether we have raw ICMP socket access at startup.
|
||||
func (f *Forwarder) checkICMPCapability() {
|
||||
ctx, cancel := context.WithTimeout(context.Background(), 100*time.Millisecond)
|
||||
defer cancel()
|
||||
|
||||
lc := net.ListenConfig{}
|
||||
conn, err := lc.ListenPacket(ctx, "ip4:icmp", "0.0.0.0")
|
||||
if err != nil {
|
||||
f.hasRawICMPAccess = false
|
||||
f.logger.Debug("forwarder: No raw ICMP socket access, will use ping binary fallback")
|
||||
return
|
||||
}
|
||||
|
||||
if err := conn.Close(); err != nil {
|
||||
f.logger.Debug1("forwarder: Failed to close ICMP capability test socket: %v", err)
|
||||
}
|
||||
|
||||
f.hasRawICMPAccess = true
|
||||
f.logger.Debug("forwarder: Raw ICMP socket access available")
|
||||
}
|
||||
|
||||
@@ -2,11 +2,8 @@ package forwarder
|
||||
|
||||
import (
|
||||
"context"
|
||||
"fmt"
|
||||
"net"
|
||||
"net/netip"
|
||||
"os/exec"
|
||||
"runtime"
|
||||
"time"
|
||||
|
||||
"github.com/google/uuid"
|
||||
@@ -17,95 +14,30 @@ import (
|
||||
)
|
||||
|
||||
// handleICMP handles ICMP packets from the network stack
|
||||
func (f *Forwarder) handleICMP(id stack.TransportEndpointID, pkt *stack.PacketBuffer) bool {
|
||||
func (f *Forwarder) handleICMP(id stack.TransportEndpointID, pkt stack.PacketBufferPtr) bool {
|
||||
icmpHdr := header.ICMPv4(pkt.TransportHeader().View().AsSlice())
|
||||
icmpType := uint8(icmpHdr.Type())
|
||||
icmpCode := uint8(icmpHdr.Code())
|
||||
|
||||
flowID := uuid.New()
|
||||
f.sendICMPEvent(nftypes.TypeStart, flowID, id, uint8(icmpHdr.Type()), uint8(icmpHdr.Code()), 0, 0)
|
||||
|
||||
// For Echo Requests, send and wait for response
|
||||
if icmpHdr.Type() == header.ICMPv4Echo {
|
||||
return f.handleICMPEcho(flowID, id, pkt, uint8(icmpHdr.Type()), uint8(icmpHdr.Code()))
|
||||
}
|
||||
|
||||
// For other ICMP types (Time Exceeded, Destination Unreachable, etc), forward without waiting
|
||||
if !f.hasRawICMPAccess {
|
||||
f.logger.Debug2("forwarder: Cannot handle ICMP type %v without raw socket access for %v", icmpHdr.Type(), epID(id))
|
||||
return false
|
||||
}
|
||||
|
||||
icmpData := stack.PayloadSince(pkt.TransportHeader()).AsSlice()
|
||||
conn, err := f.forwardICMPPacket(id, icmpData, uint8(icmpHdr.Type()), uint8(icmpHdr.Code()), 100*time.Millisecond)
|
||||
if err != nil {
|
||||
f.logger.Error2("forwarder: Failed to forward ICMP packet for %v: %v", epID(id), err)
|
||||
if header.ICMPv4Type(icmpType) == header.ICMPv4EchoReply {
|
||||
// dont process our own replies
|
||||
return true
|
||||
}
|
||||
if err := conn.Close(); err != nil {
|
||||
f.logger.Debug1("forwarder: Failed to close ICMP socket: %v", err)
|
||||
}
|
||||
|
||||
return true
|
||||
}
|
||||
flowID := uuid.New()
|
||||
f.sendICMPEvent(nftypes.TypeStart, flowID, id, icmpType, icmpCode, 0, 0)
|
||||
|
||||
// handleICMPEcho handles ICMP echo requests asynchronously with rate limiting.
|
||||
func (f *Forwarder) handleICMPEcho(flowID uuid.UUID, id stack.TransportEndpointID, pkt *stack.PacketBuffer, icmpType, icmpCode uint8) bool {
|
||||
select {
|
||||
case f.pingSemaphore <- struct{}{}:
|
||||
icmpData := stack.PayloadSince(pkt.TransportHeader()).ToSlice()
|
||||
rxBytes := pkt.Size()
|
||||
|
||||
go func() {
|
||||
defer func() { <-f.pingSemaphore }()
|
||||
|
||||
if f.hasRawICMPAccess {
|
||||
f.handleICMPViaSocket(flowID, id, icmpType, icmpCode, icmpData, rxBytes)
|
||||
} else {
|
||||
f.handleICMPViaPing(flowID, id, icmpType, icmpCode, icmpData, rxBytes)
|
||||
}
|
||||
}()
|
||||
default:
|
||||
f.logger.Debug3("forwarder: ICMP rate limit exceeded for %v type %v code %v",
|
||||
epID(id), icmpType, icmpCode)
|
||||
}
|
||||
return true
|
||||
}
|
||||
|
||||
// forwardICMPPacket creates a raw ICMP socket and sends the packet, returning the connection.
|
||||
// The caller is responsible for closing the returned connection.
|
||||
func (f *Forwarder) forwardICMPPacket(id stack.TransportEndpointID, payload []byte, icmpType, icmpCode uint8, timeout time.Duration) (net.PacketConn, error) {
|
||||
ctx, cancel := context.WithTimeout(f.ctx, timeout)
|
||||
ctx, cancel := context.WithTimeout(f.ctx, 5*time.Second)
|
||||
defer cancel()
|
||||
|
||||
lc := net.ListenConfig{}
|
||||
// TODO: support non-root
|
||||
conn, err := lc.ListenPacket(ctx, "ip4:icmp", "0.0.0.0")
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("create ICMP socket: %w", err)
|
||||
}
|
||||
f.logger.Error2("forwarder: Failed to create ICMP socket for %v: %v", epID(id), err)
|
||||
|
||||
dstIP := f.determineDialAddr(id.LocalAddress)
|
||||
dst := &net.IPAddr{IP: dstIP}
|
||||
|
||||
if _, err = conn.WriteTo(payload, dst); err != nil {
|
||||
if closeErr := conn.Close(); closeErr != nil {
|
||||
f.logger.Debug1("forwarder: Failed to close ICMP socket: %v", closeErr)
|
||||
}
|
||||
return nil, fmt.Errorf("write ICMP packet: %w", err)
|
||||
}
|
||||
|
||||
f.logger.Trace3("forwarder: Forwarded ICMP packet %v type %v code %v",
|
||||
epID(id), icmpType, icmpCode)
|
||||
|
||||
return conn, nil
|
||||
}
|
||||
|
||||
// handleICMPViaSocket handles ICMP echo requests using raw sockets.
|
||||
func (f *Forwarder) handleICMPViaSocket(flowID uuid.UUID, id stack.TransportEndpointID, icmpType, icmpCode uint8, icmpData []byte, rxBytes int) {
|
||||
sendTime := time.Now()
|
||||
|
||||
conn, err := f.forwardICMPPacket(id, icmpData, icmpType, icmpCode, 5*time.Second)
|
||||
if err != nil {
|
||||
f.logger.Error2("forwarder: Failed to send ICMP packet for %v: %v", epID(id), err)
|
||||
return
|
||||
// This will make netstack reply on behalf of the original destination, that's ok for now
|
||||
return false
|
||||
}
|
||||
defer func() {
|
||||
if err := conn.Close(); err != nil {
|
||||
@@ -113,22 +45,38 @@ func (f *Forwarder) handleICMPViaSocket(flowID uuid.UUID, id stack.TransportEndp
|
||||
}
|
||||
}()
|
||||
|
||||
txBytes := f.handleEchoResponse(conn, id)
|
||||
rtt := time.Since(sendTime).Round(10 * time.Microsecond)
|
||||
dstIP := f.determineDialAddr(id.LocalAddress)
|
||||
dst := &net.IPAddr{IP: dstIP}
|
||||
|
||||
f.logger.Trace4("forwarder: Forwarded ICMP echo reply %v type %v code %v (rtt=%v, raw socket)",
|
||||
epID(id), icmpType, icmpCode, rtt)
|
||||
fullPacket := stack.PayloadSince(pkt.TransportHeader())
|
||||
payload := fullPacket.AsSlice()
|
||||
|
||||
f.sendICMPEvent(nftypes.TypeEnd, flowID, id, icmpType, icmpCode, uint64(rxBytes), uint64(txBytes))
|
||||
if _, err = conn.WriteTo(payload, dst); err != nil {
|
||||
f.logger.Error2("forwarder: Failed to write ICMP packet for %v: %v", epID(id), err)
|
||||
return true
|
||||
}
|
||||
|
||||
f.logger.Trace3("forwarder: Forwarded ICMP packet %v type %v code %v",
|
||||
epID(id), icmpHdr.Type(), icmpHdr.Code())
|
||||
|
||||
// For Echo Requests, send and handle response
|
||||
if header.ICMPv4Type(icmpType) == header.ICMPv4Echo {
|
||||
rxBytes := pkt.Size()
|
||||
txBytes := f.handleEchoResponse(icmpHdr, conn, id)
|
||||
f.sendICMPEvent(nftypes.TypeEnd, flowID, id, icmpType, icmpCode, uint64(rxBytes), uint64(txBytes))
|
||||
}
|
||||
|
||||
// For other ICMP types (Time Exceeded, Destination Unreachable, etc) do nothing
|
||||
return true
|
||||
}
|
||||
|
||||
func (f *Forwarder) handleEchoResponse(conn net.PacketConn, id stack.TransportEndpointID) int {
|
||||
func (f *Forwarder) handleEchoResponse(icmpHdr header.ICMPv4, conn net.PacketConn, id stack.TransportEndpointID) int {
|
||||
if err := conn.SetReadDeadline(time.Now().Add(5 * time.Second)); err != nil {
|
||||
f.logger.Error1("forwarder: Failed to set read deadline for ICMP response: %v", err)
|
||||
return 0
|
||||
}
|
||||
|
||||
response := make([]byte, f.endpoint.mtu.Load())
|
||||
response := make([]byte, f.endpoint.mtu)
|
||||
n, _, err := conn.ReadFrom(response)
|
||||
if err != nil {
|
||||
if !isTimeout(err) {
|
||||
@@ -137,7 +85,31 @@ func (f *Forwarder) handleEchoResponse(conn net.PacketConn, id stack.TransportEn
|
||||
return 0
|
||||
}
|
||||
|
||||
return f.injectICMPReply(id, response[:n])
|
||||
ipHdr := make([]byte, header.IPv4MinimumSize)
|
||||
ip := header.IPv4(ipHdr)
|
||||
ip.Encode(&header.IPv4Fields{
|
||||
TotalLength: uint16(header.IPv4MinimumSize + n),
|
||||
TTL: 64,
|
||||
Protocol: uint8(header.ICMPv4ProtocolNumber),
|
||||
SrcAddr: id.LocalAddress,
|
||||
DstAddr: id.RemoteAddress,
|
||||
})
|
||||
ip.SetChecksum(^ip.CalculateChecksum())
|
||||
|
||||
fullPacket := make([]byte, 0, len(ipHdr)+n)
|
||||
fullPacket = append(fullPacket, ipHdr...)
|
||||
fullPacket = append(fullPacket, response[:n]...)
|
||||
|
||||
if err := f.InjectIncomingPacket(fullPacket); err != nil {
|
||||
f.logger.Error1("forwarder: Failed to inject ICMP response: %v", err)
|
||||
|
||||
return 0
|
||||
}
|
||||
|
||||
f.logger.Trace3("forwarder: Forwarded ICMP echo reply for %v type %v code %v",
|
||||
epID(id), icmpHdr.Type(), icmpHdr.Code())
|
||||
|
||||
return len(fullPacket)
|
||||
}
|
||||
|
||||
// sendICMPEvent stores flow events for ICMP packets
|
||||
@@ -180,95 +152,3 @@ func (f *Forwarder) sendICMPEvent(typ nftypes.Type, flowID uuid.UUID, id stack.T
|
||||
|
||||
f.flowLogger.StoreEvent(fields)
|
||||
}
|
||||
|
||||
// handleICMPViaPing handles ICMP echo requests by executing the system ping binary.
|
||||
// This is used as a fallback when raw socket access is not available.
|
||||
func (f *Forwarder) handleICMPViaPing(flowID uuid.UUID, id stack.TransportEndpointID, icmpType, icmpCode uint8, icmpData []byte, rxBytes int) {
|
||||
ctx, cancel := context.WithTimeout(f.ctx, 5*time.Second)
|
||||
defer cancel()
|
||||
|
||||
dstIP := f.determineDialAddr(id.LocalAddress)
|
||||
cmd := buildPingCommand(ctx, dstIP, 5*time.Second)
|
||||
|
||||
pingStart := time.Now()
|
||||
if err := cmd.Run(); err != nil {
|
||||
f.logger.Warn4("forwarder: Ping binary failed for %v type %v code %v: %v", epID(id),
|
||||
icmpType, icmpCode, err)
|
||||
return
|
||||
}
|
||||
rtt := time.Since(pingStart).Round(10 * time.Microsecond)
|
||||
|
||||
f.logger.Trace3("forwarder: Forwarded ICMP echo request %v type %v code %v",
|
||||
epID(id), icmpType, icmpCode)
|
||||
|
||||
txBytes := f.synthesizeEchoReply(id, icmpData)
|
||||
|
||||
f.logger.Trace4("forwarder: Forwarded ICMP echo reply %v type %v code %v (rtt=%v, ping binary)",
|
||||
epID(id), icmpType, icmpCode, rtt)
|
||||
|
||||
f.sendICMPEvent(nftypes.TypeEnd, flowID, id, icmpType, icmpCode, uint64(rxBytes), uint64(txBytes))
|
||||
}
|
||||
|
||||
// buildPingCommand creates a platform-specific ping command.
|
||||
func buildPingCommand(ctx context.Context, target net.IP, timeout time.Duration) *exec.Cmd {
|
||||
timeoutSec := int(timeout.Seconds())
|
||||
if timeoutSec < 1 {
|
||||
timeoutSec = 1
|
||||
}
|
||||
|
||||
switch runtime.GOOS {
|
||||
case "linux", "android":
|
||||
return exec.CommandContext(ctx, "ping", "-c", "1", "-W", fmt.Sprintf("%d", timeoutSec), "-q", target.String())
|
||||
case "darwin", "ios":
|
||||
return exec.CommandContext(ctx, "ping", "-c", "1", "-t", fmt.Sprintf("%d", timeoutSec), "-q", target.String())
|
||||
case "freebsd":
|
||||
return exec.CommandContext(ctx, "ping", "-c", "1", "-t", fmt.Sprintf("%d", timeoutSec), target.String())
|
||||
case "openbsd", "netbsd":
|
||||
return exec.CommandContext(ctx, "ping", "-c", "1", "-w", fmt.Sprintf("%d", timeoutSec), target.String())
|
||||
case "windows":
|
||||
return exec.CommandContext(ctx, "ping", "-n", "1", "-w", fmt.Sprintf("%d", timeoutSec*1000), target.String())
|
||||
default:
|
||||
return exec.CommandContext(ctx, "ping", "-c", "1", target.String())
|
||||
}
|
||||
}
|
||||
|
||||
// synthesizeEchoReply creates an ICMP echo reply from raw ICMP data and injects it back into the network stack.
|
||||
// Returns the size of the injected packet.
|
||||
func (f *Forwarder) synthesizeEchoReply(id stack.TransportEndpointID, icmpData []byte) int {
|
||||
replyICMP := make([]byte, len(icmpData))
|
||||
copy(replyICMP, icmpData)
|
||||
|
||||
replyICMPHdr := header.ICMPv4(replyICMP)
|
||||
replyICMPHdr.SetType(header.ICMPv4EchoReply)
|
||||
replyICMPHdr.SetChecksum(0)
|
||||
replyICMPHdr.SetChecksum(header.ICMPv4Checksum(replyICMPHdr, 0))
|
||||
|
||||
return f.injectICMPReply(id, replyICMP)
|
||||
}
|
||||
|
||||
// injectICMPReply wraps an ICMP payload in an IP header and injects it into the network stack.
|
||||
// Returns the total size of the injected packet, or 0 if injection failed.
|
||||
func (f *Forwarder) injectICMPReply(id stack.TransportEndpointID, icmpPayload []byte) int {
|
||||
ipHdr := make([]byte, header.IPv4MinimumSize)
|
||||
ip := header.IPv4(ipHdr)
|
||||
ip.Encode(&header.IPv4Fields{
|
||||
TotalLength: uint16(header.IPv4MinimumSize + len(icmpPayload)),
|
||||
TTL: 64,
|
||||
Protocol: uint8(header.ICMPv4ProtocolNumber),
|
||||
SrcAddr: id.LocalAddress,
|
||||
DstAddr: id.RemoteAddress,
|
||||
})
|
||||
ip.SetChecksum(^ip.CalculateChecksum())
|
||||
|
||||
fullPacket := make([]byte, 0, len(ipHdr)+len(icmpPayload))
|
||||
fullPacket = append(fullPacket, ipHdr...)
|
||||
fullPacket = append(fullPacket, icmpPayload...)
|
||||
|
||||
// Bypass netstack and send directly to peer to avoid looping through our ICMP handler
|
||||
if err := f.endpoint.device.CreateOutboundPacket(fullPacket, id.RemoteAddress.AsSlice()); err != nil {
|
||||
f.logger.Error1("forwarder: Failed to send ICMP reply to peer: %v", err)
|
||||
return 0
|
||||
}
|
||||
|
||||
return len(fullPacket)
|
||||
}
|
||||
|
||||
@@ -4,7 +4,6 @@ import (
|
||||
"context"
|
||||
"errors"
|
||||
"fmt"
|
||||
"io"
|
||||
"net"
|
||||
"net/netip"
|
||||
"sync"
|
||||
@@ -132,10 +131,10 @@ func (f *udpForwarder) cleanup() {
|
||||
}
|
||||
|
||||
// handleUDP is called by the UDP forwarder for new packets
|
||||
func (f *Forwarder) handleUDP(r *udp.ForwarderRequest) bool {
|
||||
func (f *Forwarder) handleUDP(r *udp.ForwarderRequest) {
|
||||
if f.ctx.Err() != nil {
|
||||
f.logger.Trace("forwarder: context done, dropping UDP packet")
|
||||
return false
|
||||
return
|
||||
}
|
||||
|
||||
id := r.ID()
|
||||
@@ -145,7 +144,7 @@ func (f *Forwarder) handleUDP(r *udp.ForwarderRequest) bool {
|
||||
f.udpForwarder.RUnlock()
|
||||
if exists {
|
||||
f.logger.Trace1("forwarder: existing UDP connection for %v", epID(id))
|
||||
return true
|
||||
return
|
||||
}
|
||||
|
||||
flowID := uuid.New()
|
||||
@@ -163,7 +162,7 @@ func (f *Forwarder) handleUDP(r *udp.ForwarderRequest) bool {
|
||||
if err != nil {
|
||||
f.logger.Debug2("forwarder: UDP dial error for %v: %v", epID(id), err)
|
||||
// TODO: Send ICMP error message
|
||||
return false
|
||||
return
|
||||
}
|
||||
|
||||
// Create wait queue for blocking syscalls
|
||||
@@ -174,10 +173,10 @@ func (f *Forwarder) handleUDP(r *udp.ForwarderRequest) bool {
|
||||
if err := outConn.Close(); err != nil {
|
||||
f.logger.Debug2("forwarder: UDP outConn close error for %v: %v", epID(id), err)
|
||||
}
|
||||
return false
|
||||
return
|
||||
}
|
||||
|
||||
inConn := gonet.NewUDPConn(&wq, ep)
|
||||
inConn := gonet.NewUDPConn(f.stack, &wq, ep)
|
||||
connCtx, connCancel := context.WithCancel(f.ctx)
|
||||
|
||||
pConn := &udpPacketConn{
|
||||
@@ -200,7 +199,7 @@ func (f *Forwarder) handleUDP(r *udp.ForwarderRequest) bool {
|
||||
if err := outConn.Close(); err != nil {
|
||||
f.logger.Debug2("forwarder: UDP outConn close error for %v: %v", epID(id), err)
|
||||
}
|
||||
return true
|
||||
return
|
||||
}
|
||||
f.udpForwarder.conns[id] = pConn
|
||||
f.udpForwarder.Unlock()
|
||||
@@ -209,7 +208,6 @@ func (f *Forwarder) handleUDP(r *udp.ForwarderRequest) bool {
|
||||
f.logger.Trace1("forwarder: established UDP connection %v", epID(id))
|
||||
|
||||
go f.proxyUDP(connCtx, pConn, id, ep)
|
||||
return true
|
||||
}
|
||||
|
||||
func (f *Forwarder) proxyUDP(ctx context.Context, pConn *udpPacketConn, id stack.TransportEndpointID, ep tcpip.Endpoint) {
|
||||
@@ -350,7 +348,7 @@ func (c *udpPacketConn) copy(ctx context.Context, dst net.Conn, src net.Conn, bu
|
||||
}
|
||||
|
||||
func isClosedError(err error) bool {
|
||||
return errors.Is(err, net.ErrClosed) || errors.Is(err, context.Canceled) || errors.Is(err, io.EOF)
|
||||
return errors.Is(err, net.ErrClosed) || errors.Is(err, context.Canceled)
|
||||
}
|
||||
|
||||
func isTimeout(err error) bool {
|
||||
|
||||
@@ -130,7 +130,6 @@ func (m *localIPManager) UpdateLocalIPs(iface common.IFaceMapper) (err error) {
|
||||
// 127.0.0.0/8
|
||||
newIPv4Bitmap[127] = &ipv4LowBitmap{}
|
||||
for i := 0; i < 8192; i++ {
|
||||
// #nosec G602 -- bitmap is defined as [8192]uint32, loop range is correct
|
||||
newIPv4Bitmap[127].bitmap[i] = 0xFFFFFFFF
|
||||
}
|
||||
|
||||
|
||||
@@ -218,7 +218,7 @@ func BenchmarkIPChecks(b *testing.B) {
|
||||
b.ResetTimer()
|
||||
for i := 0; i < b.N; i++ {
|
||||
// nolint:gosimple
|
||||
_ = mapManager.localIPs[ip.String()]
|
||||
_, _ = mapManager.localIPs[ip.String()]
|
||||
}
|
||||
})
|
||||
|
||||
@@ -227,7 +227,7 @@ func BenchmarkIPChecks(b *testing.B) {
|
||||
b.ResetTimer()
|
||||
for i := 0; i < b.N; i++ {
|
||||
// nolint:gosimple
|
||||
_ = mapManager.localIPs[ip.String()]
|
||||
_, _ = mapManager.localIPs[ip.String()]
|
||||
}
|
||||
})
|
||||
}
|
||||
|
||||
@@ -168,15 +168,6 @@ func (l *Logger) Warn3(format string, arg1, arg2, arg3 any) {
|
||||
}
|
||||
}
|
||||
|
||||
func (l *Logger) Warn4(format string, arg1, arg2, arg3, arg4 any) {
|
||||
if l.level.Load() >= uint32(LevelWarn) {
|
||||
select {
|
||||
case l.msgChannel <- logMessage{level: LevelWarn, format: format, arg1: arg1, arg2: arg2, arg3: arg3, arg4: arg4}:
|
||||
default:
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
func (l *Logger) Debug1(format string, arg1 any) {
|
||||
if l.level.Load() >= uint32(LevelDebug) {
|
||||
select {
|
||||
|
||||
@@ -234,10 +234,9 @@ func TestInboundPortDNATNegative(t *testing.T) {
|
||||
require.False(t, translated, "Packet should NOT be translated for %s", tc.name)
|
||||
|
||||
d = parsePacket(t, packet)
|
||||
switch tc.protocol {
|
||||
case layers.IPProtocolTCP:
|
||||
if tc.protocol == layers.IPProtocolTCP {
|
||||
require.Equal(t, tc.dstPort, uint16(d.tcp.DstPort), "Port should remain unchanged")
|
||||
case layers.IPProtocolUDP:
|
||||
} else if tc.protocol == layers.IPProtocolUDP {
|
||||
require.Equal(t, tc.dstPort, uint16(d.udp.DstPort), "Port should remain unchanged")
|
||||
}
|
||||
})
|
||||
|
||||
@@ -34,7 +34,7 @@ type RouteRule struct {
|
||||
sources []netip.Prefix
|
||||
dstSet firewall.Set
|
||||
destinations []netip.Prefix
|
||||
protoLayer gopacket.LayerType
|
||||
proto firewall.Protocol
|
||||
srcPort *firewall.Port
|
||||
dstPort *firewall.Port
|
||||
action firewall.Action
|
||||
|
||||
@@ -379,9 +379,9 @@ func (m *Manager) handleNativeRouter(trace *PacketTrace) *PacketTrace {
|
||||
}
|
||||
|
||||
func (m *Manager) handleRouteACLs(trace *PacketTrace, d *decoder, srcIP, dstIP netip.Addr) *PacketTrace {
|
||||
protoLayer := d.decoded[1]
|
||||
proto, _ := getProtocolFromPacket(d)
|
||||
srcPort, dstPort := getPortsFromPacket(d)
|
||||
id, allowed := m.routeACLsPass(srcIP, dstIP, protoLayer, srcPort, dstPort)
|
||||
id, allowed := m.routeACLsPass(srcIP, dstIP, proto, srcPort, dstPort)
|
||||
|
||||
strId := string(id)
|
||||
if id == nil {
|
||||
|
||||
@@ -27,23 +27,8 @@ type receiverCreator struct {
|
||||
iceBind *ICEBind
|
||||
}
|
||||
|
||||
func (rc receiverCreator) CreateReceiverFn(pc wgConn.BatchReader, conn *net.UDPConn, rxOffload bool, msgPool *sync.Pool) wgConn.ReceiveFunc {
|
||||
if ipv4PC, ok := pc.(*ipv4.PacketConn); ok {
|
||||
return rc.iceBind.createIPv4ReceiverFn(ipv4PC, conn, rxOffload, msgPool)
|
||||
}
|
||||
// IPv6 is currently not supported in the udpmux, this is a stub for compatibility with the
|
||||
// wireguard-go ReceiverCreator interface which is called for both IPv4 and IPv6.
|
||||
return func(bufs [][]byte, sizes []int, eps []wgConn.Endpoint) (n int, err error) {
|
||||
buf := bufs[0]
|
||||
size, ep, err := conn.ReadFromUDPAddrPort(buf)
|
||||
if err != nil {
|
||||
return 0, err
|
||||
}
|
||||
sizes[0] = size
|
||||
stdEp := &wgConn.StdNetEndpoint{AddrPort: ep}
|
||||
eps[0] = stdEp
|
||||
return 1, nil
|
||||
}
|
||||
func (rc receiverCreator) CreateIPv4ReceiverFn(pc *ipv4.PacketConn, conn *net.UDPConn, rxOffload bool, msgPool *sync.Pool) wgConn.ReceiveFunc {
|
||||
return rc.iceBind.createIPv4ReceiverFn(pc, conn, rxOffload, msgPool)
|
||||
}
|
||||
|
||||
// ICEBind is a bind implementation with two main features:
|
||||
|
||||
@@ -1,7 +1,9 @@
|
||||
//go:build ios
|
||||
// +build ios
|
||||
|
||||
package device
|
||||
|
||||
import (
|
||||
"fmt"
|
||||
"os"
|
||||
|
||||
log "github.com/sirupsen/logrus"
|
||||
@@ -43,31 +45,10 @@ func NewTunDevice(name string, address wgaddr.Address, port int, key string, mtu
|
||||
}
|
||||
}
|
||||
|
||||
// ErrInvalidTunnelFD is returned when the tunnel file descriptor is invalid (0).
|
||||
// This typically means the Swift code couldn't find the utun control socket.
|
||||
var ErrInvalidTunnelFD = fmt.Errorf("invalid tunnel file descriptor: fd is 0 (Swift failed to locate utun socket)")
|
||||
|
||||
func (t *TunDevice) Create() (WGConfigurer, error) {
|
||||
log.Infof("create tun interface")
|
||||
|
||||
var tunDevice tun.Device
|
||||
var err error
|
||||
|
||||
// Validate the tunnel file descriptor.
|
||||
// On iOS/tvOS, the FD must be provided by the NEPacketTunnelProvider.
|
||||
// A value of 0 means the Swift code couldn't find the utun control socket
|
||||
// (the low-level APIs like ctl_info, sockaddr_ctl may not be exposed in
|
||||
// tvOS SDK headers). This is a hard error - there's no viable fallback
|
||||
// since tun.CreateTUN() cannot work within the iOS/tvOS sandbox.
|
||||
if t.tunFd == 0 {
|
||||
log.Errorf("Tunnel file descriptor is 0 - Swift code failed to locate the utun control socket. " +
|
||||
"On tvOS, ensure the NEPacketTunnelProvider is properly configured and the tunnel is started.")
|
||||
return nil, ErrInvalidTunnelFD
|
||||
}
|
||||
|
||||
// Normal iOS/tvOS path: use the provided file descriptor from NEPacketTunnelProvider
|
||||
var dupTunFd int
|
||||
dupTunFd, err = unix.Dup(t.tunFd)
|
||||
dupTunFd, err := unix.Dup(t.tunFd)
|
||||
if err != nil {
|
||||
log.Errorf("Unable to dup tun fd: %v", err)
|
||||
return nil, err
|
||||
@@ -79,7 +60,7 @@ func (t *TunDevice) Create() (WGConfigurer, error) {
|
||||
_ = unix.Close(dupTunFd)
|
||||
return nil, err
|
||||
}
|
||||
tunDevice, err = tun.CreateTUNFromFile(os.NewFile(uintptr(dupTunFd), "/dev/tun"), 0)
|
||||
tunDevice, err := tun.CreateTUNFromFile(os.NewFile(uintptr(dupTunFd), "/dev/tun"), 0)
|
||||
if err != nil {
|
||||
log.Errorf("Unable to create new tun device from fd: %v", err)
|
||||
_ = unix.Close(dupTunFd)
|
||||
|
||||
@@ -3,19 +3,12 @@
|
||||
package wgproxy
|
||||
|
||||
import (
|
||||
"os"
|
||||
"strconv"
|
||||
|
||||
log "github.com/sirupsen/logrus"
|
||||
|
||||
"github.com/netbirdio/netbird/client/iface/wgproxy/ebpf"
|
||||
udpProxy "github.com/netbirdio/netbird/client/iface/wgproxy/udp"
|
||||
)
|
||||
|
||||
const (
|
||||
envDisableEBPFWGProxy = "NB_DISABLE_EBPF_WG_PROXY"
|
||||
)
|
||||
|
||||
type KernelFactory struct {
|
||||
wgPort int
|
||||
mtu uint16
|
||||
@@ -29,12 +22,6 @@ func NewKernelFactory(wgPort int, mtu uint16) *KernelFactory {
|
||||
mtu: mtu,
|
||||
}
|
||||
|
||||
if isEBPFDisabled() {
|
||||
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 {
|
||||
log.Infof("WireGuard Proxy Factory will produce UDP proxy")
|
||||
@@ -60,16 +47,3 @@ func (w *KernelFactory) Free() error {
|
||||
}
|
||||
return w.ebpfProxy.Free()
|
||||
}
|
||||
|
||||
func isEBPFDisabled() bool {
|
||||
val := os.Getenv(envDisableEBPFWGProxy)
|
||||
if val == "" {
|
||||
return false
|
||||
}
|
||||
disabled, err := strconv.ParseBool(val)
|
||||
if err != nil {
|
||||
log.Warnf("failed to parse %s: %v", envDisableEBPFWGProxy, err)
|
||||
return false
|
||||
}
|
||||
return disabled
|
||||
}
|
||||
|
||||
@@ -24,14 +24,10 @@ import (
|
||||
"github.com/netbirdio/netbird/client/internal/listener"
|
||||
"github.com/netbirdio/netbird/client/internal/peer"
|
||||
"github.com/netbirdio/netbird/client/internal/profilemanager"
|
||||
"github.com/netbirdio/netbird/client/internal/statemanager"
|
||||
"github.com/netbirdio/netbird/client/internal/stdnet"
|
||||
"github.com/netbirdio/netbird/client/internal/updatemanager"
|
||||
"github.com/netbirdio/netbird/client/internal/updatemanager/installer"
|
||||
nbnet "github.com/netbirdio/netbird/client/net"
|
||||
cProto "github.com/netbirdio/netbird/client/proto"
|
||||
"github.com/netbirdio/netbird/client/ssh"
|
||||
sshconfig "github.com/netbirdio/netbird/client/ssh/config"
|
||||
"github.com/netbirdio/netbird/client/system"
|
||||
mgm "github.com/netbirdio/netbird/shared/management/client"
|
||||
mgmProto "github.com/netbirdio/netbird/shared/management/proto"
|
||||
@@ -43,13 +39,11 @@ import (
|
||||
)
|
||||
|
||||
type ConnectClient struct {
|
||||
ctx context.Context
|
||||
config *profilemanager.Config
|
||||
statusRecorder *peer.Status
|
||||
doInitialAutoUpdate bool
|
||||
|
||||
engine *Engine
|
||||
engineMutex sync.Mutex
|
||||
ctx context.Context
|
||||
config *profilemanager.Config
|
||||
statusRecorder *peer.Status
|
||||
engine *Engine
|
||||
engineMutex sync.Mutex
|
||||
|
||||
persistSyncResponse bool
|
||||
}
|
||||
@@ -58,15 +52,13 @@ func NewConnectClient(
|
||||
ctx context.Context,
|
||||
config *profilemanager.Config,
|
||||
statusRecorder *peer.Status,
|
||||
doInitalAutoUpdate bool,
|
||||
|
||||
) *ConnectClient {
|
||||
return &ConnectClient{
|
||||
ctx: ctx,
|
||||
config: config,
|
||||
statusRecorder: statusRecorder,
|
||||
doInitialAutoUpdate: doInitalAutoUpdate,
|
||||
engineMutex: sync.Mutex{},
|
||||
ctx: ctx,
|
||||
config: config,
|
||||
statusRecorder: statusRecorder,
|
||||
engineMutex: sync.Mutex{},
|
||||
}
|
||||
}
|
||||
|
||||
@@ -170,33 +162,6 @@ func (c *ConnectClient) run(mobileDependency MobileDependency, runningChan chan
|
||||
return err
|
||||
}
|
||||
|
||||
var path string
|
||||
if runtime.GOOS == "ios" || runtime.GOOS == "android" {
|
||||
// On mobile, use the provided state file path directly
|
||||
if !fileExists(mobileDependency.StateFilePath) {
|
||||
if err := createFile(mobileDependency.StateFilePath); err != nil {
|
||||
log.Errorf("failed to create state file: %v", err)
|
||||
// we are not exiting as we can run without the state manager
|
||||
}
|
||||
}
|
||||
path = mobileDependency.StateFilePath
|
||||
} else {
|
||||
sm := profilemanager.NewServiceManager("")
|
||||
path = sm.GetStatePath()
|
||||
}
|
||||
stateManager := statemanager.New(path)
|
||||
stateManager.RegisterState(&sshconfig.ShutdownState{})
|
||||
|
||||
updateManager, err := updatemanager.NewManager(c.statusRecorder, stateManager)
|
||||
if err == nil {
|
||||
updateManager.CheckUpdateSuccess(c.ctx)
|
||||
|
||||
inst := installer.New()
|
||||
if err := inst.CleanUpInstallerFiles(); err != nil {
|
||||
log.Errorf("failed to clean up temporary installer file: %v", err)
|
||||
}
|
||||
}
|
||||
|
||||
defer c.statusRecorder.ClientStop()
|
||||
operation := func() error {
|
||||
// if context cancelled we not start new backoff cycle
|
||||
@@ -308,7 +273,7 @@ func (c *ConnectClient) run(mobileDependency MobileDependency, runningChan chan
|
||||
checks := loginResp.GetChecks()
|
||||
|
||||
c.engineMutex.Lock()
|
||||
engine := NewEngine(engineCtx, cancel, signalClient, mgmClient, relayManager, engineConfig, mobileDependency, c.statusRecorder, checks, stateManager)
|
||||
engine := NewEngine(engineCtx, cancel, signalClient, mgmClient, relayManager, engineConfig, mobileDependency, c.statusRecorder, checks)
|
||||
engine.SetSyncResponsePersistence(c.persistSyncResponse)
|
||||
c.engine = engine
|
||||
c.engineMutex.Unlock()
|
||||
@@ -318,15 +283,6 @@ func (c *ConnectClient) run(mobileDependency MobileDependency, runningChan chan
|
||||
return wrapErr(err)
|
||||
}
|
||||
|
||||
if loginResp.PeerConfig != nil && loginResp.PeerConfig.AutoUpdate != nil {
|
||||
// AutoUpdate will be true when the user click on "Connect" menu on the UI
|
||||
if c.doInitialAutoUpdate {
|
||||
log.Infof("start engine by ui, run auto-update check")
|
||||
c.engine.InitialUpdateHandling(loginResp.PeerConfig.AutoUpdate)
|
||||
c.doInitialAutoUpdate = false
|
||||
}
|
||||
}
|
||||
|
||||
log.Infof("Netbird engine started, the IP is: %s", peerConfig.GetAddress())
|
||||
state.Set(StatusConnected)
|
||||
|
||||
@@ -420,19 +376,6 @@ func (c *ConnectClient) GetLatestSyncResponse() (*mgmProto.SyncResponse, error)
|
||||
return syncResponse, nil
|
||||
}
|
||||
|
||||
// SetLogLevel sets the log level for the firewall manager if the engine is running.
|
||||
func (c *ConnectClient) SetLogLevel(level log.Level) {
|
||||
engine := c.Engine()
|
||||
if engine == nil {
|
||||
return
|
||||
}
|
||||
|
||||
fwManager := engine.GetFirewallManager()
|
||||
if fwManager != nil {
|
||||
fwManager.SetLogLevel(level)
|
||||
}
|
||||
}
|
||||
|
||||
// Status returns the current client status
|
||||
func (c *ConnectClient) Status() StatusType {
|
||||
if c == nil {
|
||||
|
||||
@@ -27,7 +27,6 @@ import (
|
||||
"github.com/netbirdio/netbird/client/anonymize"
|
||||
"github.com/netbirdio/netbird/client/internal/peer"
|
||||
"github.com/netbirdio/netbird/client/internal/profilemanager"
|
||||
"github.com/netbirdio/netbird/client/internal/updatemanager/installer"
|
||||
mgmProto "github.com/netbirdio/netbird/shared/management/proto"
|
||||
"github.com/netbirdio/netbird/util"
|
||||
)
|
||||
@@ -57,7 +56,6 @@ block.prof: Block profiling information.
|
||||
heap.prof: Heap profiling information (snapshot of memory allocations).
|
||||
allocs.prof: Allocations profiling information.
|
||||
threadcreate.prof: Thread creation profiling information.
|
||||
stack_trace.txt: Complete stack traces of all goroutines at the time of bundle creation.
|
||||
|
||||
|
||||
Anonymization Process
|
||||
@@ -111,9 +109,6 @@ go tool pprof -http=:8088 heap.prof
|
||||
|
||||
This will open a web browser tab with the profiling information.
|
||||
|
||||
Stack Trace
|
||||
The stack_trace.txt file contains a complete snapshot of all goroutine stack traces at the time the debug bundle was created.
|
||||
|
||||
Routes
|
||||
The routes.txt file contains detailed routing table information in a tabular format:
|
||||
|
||||
@@ -332,10 +327,6 @@ func (g *BundleGenerator) createArchive() error {
|
||||
log.Errorf("failed to add profiles to debug bundle: %v", err)
|
||||
}
|
||||
|
||||
if err := g.addStackTrace(); err != nil {
|
||||
log.Errorf("failed to add stack trace to debug bundle: %v", err)
|
||||
}
|
||||
|
||||
if err := g.addSyncResponse(); err != nil {
|
||||
return fmt.Errorf("add sync response: %w", err)
|
||||
}
|
||||
@@ -363,10 +354,6 @@ func (g *BundleGenerator) createArchive() error {
|
||||
log.Errorf("failed to add systemd logs: %v", err)
|
||||
}
|
||||
|
||||
if err := g.addUpdateLogs(); err != nil {
|
||||
log.Errorf("failed to add updater logs: %v", err)
|
||||
}
|
||||
|
||||
return nil
|
||||
}
|
||||
|
||||
@@ -535,18 +522,6 @@ func (g *BundleGenerator) addProf() (err error) {
|
||||
return nil
|
||||
}
|
||||
|
||||
func (g *BundleGenerator) addStackTrace() error {
|
||||
buf := make([]byte, 5242880) // 5 MB buffer
|
||||
n := runtime.Stack(buf, true)
|
||||
|
||||
stackTrace := bytes.NewReader(buf[:n])
|
||||
if err := g.addFileToZip(stackTrace, "stack_trace.txt"); err != nil {
|
||||
return fmt.Errorf("add stack trace file to zip: %w", err)
|
||||
}
|
||||
|
||||
return nil
|
||||
}
|
||||
|
||||
func (g *BundleGenerator) addInterfaces() error {
|
||||
interfaces, err := net.Interfaces()
|
||||
if err != nil {
|
||||
@@ -655,29 +630,6 @@ func (g *BundleGenerator) addStateFile() error {
|
||||
return nil
|
||||
}
|
||||
|
||||
func (g *BundleGenerator) addUpdateLogs() error {
|
||||
inst := installer.New()
|
||||
logFiles := inst.LogFiles()
|
||||
if len(logFiles) == 0 {
|
||||
return nil
|
||||
}
|
||||
|
||||
log.Infof("adding updater logs")
|
||||
for _, logFile := range logFiles {
|
||||
data, err := os.ReadFile(logFile)
|
||||
if err != nil {
|
||||
log.Warnf("failed to read update log file %s: %v", logFile, err)
|
||||
continue
|
||||
}
|
||||
|
||||
baseName := filepath.Base(logFile)
|
||||
if err := g.addFileToZip(bytes.NewReader(data), filepath.Join("update-logs", baseName)); err != nil {
|
||||
return fmt.Errorf("add update log file %s to zip: %w", baseName, err)
|
||||
}
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
func (g *BundleGenerator) addCorruptedStateFiles() error {
|
||||
sm := profilemanager.NewServiceManager("")
|
||||
pattern := sm.GetStatePath()
|
||||
|
||||
@@ -507,13 +507,15 @@ func formatPayloadWithCmp(p *expr.Payload, cmp *expr.Cmp) string {
|
||||
if p.Base == expr.PayloadBaseNetworkHeader {
|
||||
switch p.Offset {
|
||||
case 12:
|
||||
switch p.Len {
|
||||
case 4, 2:
|
||||
if p.Len == 4 {
|
||||
return fmt.Sprintf("ip saddr %s %s", formatCmpOp(cmp.Op), formatIPBytes(cmp.Data))
|
||||
} else if p.Len == 2 {
|
||||
return fmt.Sprintf("ip saddr %s %s", formatCmpOp(cmp.Op), formatIPBytes(cmp.Data))
|
||||
}
|
||||
case 16:
|
||||
switch p.Len {
|
||||
case 4, 2:
|
||||
if p.Len == 4 {
|
||||
return fmt.Sprintf("ip daddr %s %s", formatCmpOp(cmp.Op), formatIPBytes(cmp.Data))
|
||||
} else if p.Len == 2 {
|
||||
return fmt.Sprintf("ip daddr %s %s", formatCmpOp(cmp.Op), formatIPBytes(cmp.Data))
|
||||
}
|
||||
}
|
||||
|
||||
@@ -76,7 +76,7 @@ func collectPTRRecords(config *nbdns.Config, prefix netip.Prefix) []nbdns.Simple
|
||||
var records []nbdns.SimpleRecord
|
||||
|
||||
for _, zone := range config.CustomZones {
|
||||
if zone.NonAuthoritative {
|
||||
if zone.SkipPTRProcess {
|
||||
continue
|
||||
}
|
||||
for _, record := range zone.Records {
|
||||
|
||||
@@ -3,21 +3,17 @@ package dns
|
||||
import (
|
||||
"fmt"
|
||||
"slices"
|
||||
"strconv"
|
||||
"strings"
|
||||
"sync"
|
||||
"time"
|
||||
|
||||
"github.com/miekg/dns"
|
||||
log "github.com/sirupsen/logrus"
|
||||
|
||||
"github.com/netbirdio/netbird/client/internal/dns/resutil"
|
||||
)
|
||||
|
||||
const (
|
||||
PriorityMgmtCache = 150
|
||||
PriorityDNSRoute = 100
|
||||
PriorityLocal = 75
|
||||
PriorityLocal = 100
|
||||
PriorityDNSRoute = 75
|
||||
PriorityUpstream = 50
|
||||
PriorityDefault = 1
|
||||
PriorityFallback = -100
|
||||
@@ -47,23 +43,7 @@ type HandlerChain struct {
|
||||
type ResponseWriterChain struct {
|
||||
dns.ResponseWriter
|
||||
origPattern string
|
||||
requestID string
|
||||
shouldContinue bool
|
||||
response *dns.Msg
|
||||
meta map[string]string
|
||||
}
|
||||
|
||||
// RequestID returns the request ID for tracing
|
||||
func (w *ResponseWriterChain) RequestID() string {
|
||||
return w.requestID
|
||||
}
|
||||
|
||||
// SetMeta sets a metadata key-value pair for logging
|
||||
func (w *ResponseWriterChain) SetMeta(key, value string) {
|
||||
if w.meta == nil {
|
||||
w.meta = make(map[string]string)
|
||||
}
|
||||
w.meta[key] = value
|
||||
}
|
||||
|
||||
func (w *ResponseWriterChain) WriteMsg(m *dns.Msg) error {
|
||||
@@ -72,7 +52,6 @@ func (w *ResponseWriterChain) WriteMsg(m *dns.Msg) error {
|
||||
w.shouldContinue = true
|
||||
return nil
|
||||
}
|
||||
w.response = m
|
||||
return w.ResponseWriter.WriteMsg(m)
|
||||
}
|
||||
|
||||
@@ -122,8 +101,6 @@ func (c *HandlerChain) AddHandler(pattern string, handler dns.Handler, priority
|
||||
|
||||
pos := c.findHandlerPosition(entry)
|
||||
c.handlers = append(c.handlers[:pos], append([]HandlerEntry{entry}, c.handlers[pos:]...)...)
|
||||
|
||||
c.logHandlers()
|
||||
}
|
||||
|
||||
// findHandlerPosition determines where to insert a new handler based on priority and specificity
|
||||
@@ -163,109 +140,68 @@ func (c *HandlerChain) removeEntry(pattern string, priority int) {
|
||||
for i := len(c.handlers) - 1; i >= 0; i-- {
|
||||
entry := c.handlers[i]
|
||||
if strings.EqualFold(entry.OrigPattern, pattern) && entry.Priority == priority {
|
||||
log.Debugf("removing handler pattern: domain=%s priority=%d", entry.OrigPattern, priority)
|
||||
c.handlers = append(c.handlers[:i], c.handlers[i+1:]...)
|
||||
c.logHandlers()
|
||||
break
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// logHandlers logs the current handler chain state. Caller must hold the lock.
|
||||
func (c *HandlerChain) logHandlers() {
|
||||
if !log.IsLevelEnabled(log.TraceLevel) {
|
||||
return
|
||||
}
|
||||
|
||||
var b strings.Builder
|
||||
b.WriteString("handler chain (" + strconv.Itoa(len(c.handlers)) + "):\n")
|
||||
for _, h := range c.handlers {
|
||||
b.WriteString(" - pattern: domain=" + h.Pattern + " original: domain=" + h.OrigPattern +
|
||||
" wildcard=" + strconv.FormatBool(h.IsWildcard) +
|
||||
" match_subdomain=" + strconv.FormatBool(h.MatchSubdomains) +
|
||||
" priority=" + strconv.Itoa(h.Priority) + "\n")
|
||||
}
|
||||
log.Trace(strings.TrimSuffix(b.String(), "\n"))
|
||||
}
|
||||
|
||||
func (c *HandlerChain) ServeDNS(w dns.ResponseWriter, r *dns.Msg) {
|
||||
if len(r.Question) == 0 {
|
||||
return
|
||||
}
|
||||
|
||||
startTime := time.Now()
|
||||
requestID := resutil.GenerateRequestID()
|
||||
logger := log.WithFields(log.Fields{
|
||||
"request_id": requestID,
|
||||
"dns_id": fmt.Sprintf("%04x", r.Id),
|
||||
})
|
||||
|
||||
question := r.Question[0]
|
||||
qname := strings.ToLower(question.Name)
|
||||
qname := strings.ToLower(r.Question[0].Name)
|
||||
|
||||
c.mu.RLock()
|
||||
handlers := slices.Clone(c.handlers)
|
||||
c.mu.RUnlock()
|
||||
|
||||
if log.IsLevelEnabled(log.TraceLevel) {
|
||||
var b strings.Builder
|
||||
b.WriteString(fmt.Sprintf("DNS request domain=%s, handlers (%d):\n", qname, len(handlers)))
|
||||
for _, h := range handlers {
|
||||
b.WriteString(fmt.Sprintf(" - pattern: domain=%s original: domain=%s wildcard=%v match_subdomain=%v priority=%d\n",
|
||||
h.Pattern, h.OrigPattern, h.IsWildcard, h.MatchSubdomains, h.Priority))
|
||||
}
|
||||
log.Trace(strings.TrimSuffix(b.String(), "\n"))
|
||||
}
|
||||
|
||||
// Try handlers in priority order
|
||||
for _, entry := range handlers {
|
||||
if !c.isHandlerMatch(qname, entry) {
|
||||
continue
|
||||
}
|
||||
matched := c.isHandlerMatch(qname, entry)
|
||||
|
||||
handlerName := entry.OrigPattern
|
||||
if s, ok := entry.Handler.(interface{ String() string }); ok {
|
||||
handlerName = s.String()
|
||||
}
|
||||
if matched {
|
||||
log.Tracef("handler matched: domain=%s -> pattern=%s wildcard=%v match_subdomain=%v priority=%d",
|
||||
qname, entry.OrigPattern, entry.IsWildcard, entry.MatchSubdomains, entry.Priority)
|
||||
|
||||
logger.Tracef("question: domain=%s type=%s class=%s -> handler=%s pattern=%s wildcard=%v match_subdomain=%v priority=%d",
|
||||
qname, dns.TypeToString[question.Qtype], dns.ClassToString[question.Qclass],
|
||||
handlerName, entry.OrigPattern, entry.IsWildcard, entry.MatchSubdomains, entry.Priority)
|
||||
|
||||
chainWriter := &ResponseWriterChain{
|
||||
ResponseWriter: w,
|
||||
origPattern: entry.OrigPattern,
|
||||
requestID: requestID,
|
||||
}
|
||||
entry.Handler.ServeDNS(chainWriter, r)
|
||||
|
||||
// If handler wants to continue, try next handler
|
||||
if chainWriter.shouldContinue {
|
||||
if entry.Priority != PriorityMgmtCache {
|
||||
logger.Tracef("handler requested continue for domain=%s", qname)
|
||||
chainWriter := &ResponseWriterChain{
|
||||
ResponseWriter: w,
|
||||
origPattern: entry.OrigPattern,
|
||||
}
|
||||
continue
|
||||
}
|
||||
entry.Handler.ServeDNS(chainWriter, r)
|
||||
|
||||
c.logResponse(logger, chainWriter, qname, startTime)
|
||||
return
|
||||
// If handler wants to continue, try next handler
|
||||
if chainWriter.shouldContinue {
|
||||
// Only log continue for non-management cache handlers to reduce noise
|
||||
if entry.Priority != PriorityMgmtCache {
|
||||
log.Tracef("handler requested continue to next handler for domain=%s", qname)
|
||||
}
|
||||
continue
|
||||
}
|
||||
return
|
||||
}
|
||||
}
|
||||
|
||||
// No handler matched or all handlers passed
|
||||
logger.Tracef("no handler found for domain=%s type=%s class=%s",
|
||||
qname, dns.TypeToString[question.Qtype], dns.ClassToString[question.Qclass])
|
||||
log.Tracef("no handler found for domain=%s", qname)
|
||||
resp := &dns.Msg{}
|
||||
resp.SetRcode(r, dns.RcodeRefused)
|
||||
if err := w.WriteMsg(resp); err != nil {
|
||||
logger.Errorf("failed to write DNS response: %v", err)
|
||||
log.Errorf("failed to write DNS response: %v", err)
|
||||
}
|
||||
}
|
||||
|
||||
func (c *HandlerChain) logResponse(logger *log.Entry, cw *ResponseWriterChain, qname string, startTime time.Time) {
|
||||
if cw.response == nil {
|
||||
return
|
||||
}
|
||||
|
||||
var meta string
|
||||
for k, v := range cw.meta {
|
||||
meta += " " + k + "=" + v
|
||||
}
|
||||
|
||||
logger.Tracef("response: domain=%s rcode=%s answers=%s%s took=%s",
|
||||
qname, dns.RcodeToString[cw.response.Rcode], resutil.FormatAnswers(cw.response.Answer),
|
||||
meta, time.Since(startTime))
|
||||
}
|
||||
|
||||
func (c *HandlerChain) isHandlerMatch(qname string, entry HandlerEntry) bool {
|
||||
switch {
|
||||
case entry.Pattern == ".":
|
||||
|
||||
@@ -1,52 +1,30 @@
|
||||
package local
|
||||
|
||||
import (
|
||||
"context"
|
||||
"errors"
|
||||
"fmt"
|
||||
"net"
|
||||
"net/netip"
|
||||
"slices"
|
||||
"strings"
|
||||
"sync"
|
||||
"time"
|
||||
|
||||
"github.com/miekg/dns"
|
||||
log "github.com/sirupsen/logrus"
|
||||
"golang.org/x/exp/maps"
|
||||
|
||||
"github.com/netbirdio/netbird/client/internal/dns/resutil"
|
||||
"github.com/netbirdio/netbird/client/internal/dns/types"
|
||||
nbdns "github.com/netbirdio/netbird/dns"
|
||||
"github.com/netbirdio/netbird/shared/management/domain"
|
||||
)
|
||||
|
||||
const externalResolutionTimeout = 4 * time.Second
|
||||
|
||||
type resolver interface {
|
||||
LookupNetIP(ctx context.Context, network, host string) ([]netip.Addr, error)
|
||||
}
|
||||
|
||||
type Resolver struct {
|
||||
mu sync.RWMutex
|
||||
records map[dns.Question][]dns.RR
|
||||
domains map[domain.Domain]struct{}
|
||||
// zones maps zone domain -> NonAuthoritative (true = non-authoritative, user-created zone)
|
||||
zones map[domain.Domain]bool
|
||||
resolver resolver
|
||||
|
||||
ctx context.Context
|
||||
cancel context.CancelFunc
|
||||
}
|
||||
|
||||
func NewResolver() *Resolver {
|
||||
ctx, cancel := context.WithCancel(context.Background())
|
||||
return &Resolver{
|
||||
records: make(map[dns.Question][]dns.RR),
|
||||
domains: make(map[domain.Domain]struct{}),
|
||||
zones: make(map[domain.Domain]bool),
|
||||
ctx: ctx,
|
||||
cancel: cancel,
|
||||
}
|
||||
}
|
||||
|
||||
@@ -59,18 +37,7 @@ func (d *Resolver) String() string {
|
||||
return fmt.Sprintf("LocalResolver [%d records]", len(d.records))
|
||||
}
|
||||
|
||||
func (d *Resolver) Stop() {
|
||||
if d.cancel != nil {
|
||||
d.cancel()
|
||||
}
|
||||
|
||||
d.mu.Lock()
|
||||
defer d.mu.Unlock()
|
||||
|
||||
maps.Clear(d.records)
|
||||
maps.Clear(d.domains)
|
||||
maps.Clear(d.zones)
|
||||
}
|
||||
func (d *Resolver) Stop() {}
|
||||
|
||||
// ID returns the unique handler ID
|
||||
func (d *Resolver) ID() types.HandlerID {
|
||||
@@ -81,85 +48,35 @@ func (d *Resolver) ProbeAvailability() {}
|
||||
|
||||
// ServeDNS handles a DNS request
|
||||
func (d *Resolver) ServeDNS(w dns.ResponseWriter, r *dns.Msg) {
|
||||
logger := log.WithField("request_id", resutil.GetRequestID(w))
|
||||
|
||||
if len(r.Question) == 0 {
|
||||
logger.Debug("received local resolver request with no question")
|
||||
log.Debugf("received local resolver request with no question")
|
||||
return
|
||||
}
|
||||
question := r.Question[0]
|
||||
question.Name = strings.ToLower(dns.Fqdn(question.Name))
|
||||
|
||||
log.Tracef("received local question: domain=%s type=%v class=%v", r.Question[0].Name, question.Qtype, question.Qclass)
|
||||
|
||||
replyMessage := &dns.Msg{}
|
||||
replyMessage.SetReply(r)
|
||||
replyMessage.RecursionAvailable = true
|
||||
|
||||
result := d.lookupRecords(logger, question)
|
||||
replyMessage.Authoritative = !result.hasExternalData
|
||||
replyMessage.Answer = result.records
|
||||
replyMessage.Rcode = d.determineRcode(question, result)
|
||||
|
||||
if replyMessage.Rcode == dns.RcodeNameError && d.shouldFallthrough(question.Name) {
|
||||
d.continueToNext(logger, w, r)
|
||||
return
|
||||
// lookup all records matching the question
|
||||
records := d.lookupRecords(question)
|
||||
if len(records) > 0 {
|
||||
replyMessage.Rcode = dns.RcodeSuccess
|
||||
replyMessage.Answer = append(replyMessage.Answer, records...)
|
||||
} else {
|
||||
// Check if we have any records for this domain name with different types
|
||||
if d.hasRecordsForDomain(domain.Domain(question.Name)) {
|
||||
replyMessage.Rcode = dns.RcodeSuccess // NOERROR with 0 records
|
||||
} else {
|
||||
replyMessage.Rcode = dns.RcodeNameError // NXDOMAIN
|
||||
}
|
||||
}
|
||||
|
||||
if err := w.WriteMsg(replyMessage); err != nil {
|
||||
logger.Warnf("failed to write the local resolver response: %v", err)
|
||||
}
|
||||
}
|
||||
|
||||
// determineRcode returns the appropriate DNS response code.
|
||||
// Per RFC 6604, CNAME chains should return the rcode of the final target resolution,
|
||||
// even if CNAME records are included in the answer.
|
||||
func (d *Resolver) determineRcode(question dns.Question, result lookupResult) int {
|
||||
// Use the rcode from lookup - this properly handles CNAME chains where
|
||||
// the target may be NXDOMAIN or SERVFAIL even though we have CNAME records
|
||||
if result.rcode != 0 {
|
||||
return result.rcode
|
||||
}
|
||||
|
||||
// No records found, but domain exists with different record types (NODATA)
|
||||
if d.hasRecordsForDomain(domain.Domain(question.Name)) {
|
||||
return dns.RcodeSuccess
|
||||
}
|
||||
|
||||
return dns.RcodeNameError
|
||||
}
|
||||
|
||||
// findZone finds the matching zone for a query name using reverse suffix lookup.
|
||||
// Returns (nonAuthoritative, found). This is O(k) where k = number of labels in qname.
|
||||
func (d *Resolver) findZone(qname string) (nonAuthoritative bool, found bool) {
|
||||
qname = strings.ToLower(dns.Fqdn(qname))
|
||||
for {
|
||||
if nonAuth, ok := d.zones[domain.Domain(qname)]; ok {
|
||||
return nonAuth, true
|
||||
}
|
||||
// Move to parent domain
|
||||
idx := strings.Index(qname, ".")
|
||||
if idx == -1 || idx == len(qname)-1 {
|
||||
return false, false
|
||||
}
|
||||
qname = qname[idx+1:]
|
||||
}
|
||||
}
|
||||
|
||||
// shouldFallthrough checks if the query should fallthrough to the next handler.
|
||||
// Returns true if the queried name belongs to a non-authoritative zone.
|
||||
func (d *Resolver) shouldFallthrough(qname string) bool {
|
||||
d.mu.RLock()
|
||||
defer d.mu.RUnlock()
|
||||
|
||||
nonAuth, found := d.findZone(qname)
|
||||
return found && nonAuth
|
||||
}
|
||||
|
||||
func (d *Resolver) continueToNext(logger *log.Entry, w dns.ResponseWriter, r *dns.Msg) {
|
||||
resp := &dns.Msg{}
|
||||
resp.SetRcode(r, dns.RcodeNameError)
|
||||
resp.MsgHdr.Zero = true
|
||||
if err := w.WriteMsg(resp); err != nil {
|
||||
logger.Warnf("failed to write continue signal: %v", err)
|
||||
log.Warnf("failed to write the local resolver response: %v", err)
|
||||
}
|
||||
}
|
||||
|
||||
@@ -172,27 +89,8 @@ func (d *Resolver) hasRecordsForDomain(domainName domain.Domain) bool {
|
||||
return exists
|
||||
}
|
||||
|
||||
// isInManagedZone checks if the given name falls within any of our managed zones.
|
||||
// This is used to avoid unnecessary external resolution for CNAME targets that
|
||||
// are within zones we manage - if we don't have a record for it, it doesn't exist.
|
||||
// Caller must NOT hold the lock.
|
||||
func (d *Resolver) isInManagedZone(name string) bool {
|
||||
d.mu.RLock()
|
||||
defer d.mu.RUnlock()
|
||||
|
||||
_, found := d.findZone(name)
|
||||
return found
|
||||
}
|
||||
|
||||
// lookupResult contains the result of a DNS lookup operation.
|
||||
type lookupResult struct {
|
||||
records []dns.RR
|
||||
rcode int
|
||||
hasExternalData bool
|
||||
}
|
||||
|
||||
// lookupRecords fetches *all* DNS records matching the first question in r.
|
||||
func (d *Resolver) lookupRecords(logger *log.Entry, question dns.Question) lookupResult {
|
||||
func (d *Resolver) lookupRecords(question dns.Question) []dns.RR {
|
||||
d.mu.RLock()
|
||||
records, found := d.records[question]
|
||||
|
||||
@@ -200,14 +98,10 @@ func (d *Resolver) lookupRecords(logger *log.Entry, question dns.Question) looku
|
||||
d.mu.RUnlock()
|
||||
// alternatively check if we have a cname
|
||||
if question.Qtype != dns.TypeCNAME {
|
||||
cnameQuestion := dns.Question{
|
||||
Name: question.Name,
|
||||
Qtype: dns.TypeCNAME,
|
||||
Qclass: question.Qclass,
|
||||
}
|
||||
return d.lookupCNAMEChain(logger, cnameQuestion, question.Qtype)
|
||||
question.Qtype = dns.TypeCNAME
|
||||
return d.lookupRecords(question)
|
||||
}
|
||||
return lookupResult{rcode: dns.RcodeNameError}
|
||||
return nil
|
||||
}
|
||||
|
||||
recordsCopy := slices.Clone(records)
|
||||
@@ -225,178 +119,20 @@ func (d *Resolver) lookupRecords(logger *log.Entry, question dns.Question) looku
|
||||
d.mu.Unlock()
|
||||
}
|
||||
|
||||
return lookupResult{records: recordsCopy, rcode: dns.RcodeSuccess}
|
||||
return recordsCopy
|
||||
}
|
||||
|
||||
// lookupCNAMEChain follows a CNAME chain and returns the CNAME records along with
|
||||
// the final resolved record of the requested type. This is required for musl libc
|
||||
// compatibility, which expects the full answer chain rather than just the CNAME.
|
||||
func (d *Resolver) lookupCNAMEChain(logger *log.Entry, cnameQuestion dns.Question, targetType uint16) lookupResult {
|
||||
const maxDepth = 8
|
||||
var chain []dns.RR
|
||||
|
||||
for range maxDepth {
|
||||
cnameRecords := d.getRecords(cnameQuestion)
|
||||
if len(cnameRecords) == 0 {
|
||||
break
|
||||
}
|
||||
|
||||
chain = append(chain, cnameRecords...)
|
||||
|
||||
cname, ok := cnameRecords[0].(*dns.CNAME)
|
||||
if !ok {
|
||||
break
|
||||
}
|
||||
|
||||
targetName := strings.ToLower(cname.Target)
|
||||
targetResult := d.resolveCNAMETarget(logger, targetName, targetType, cnameQuestion.Qclass)
|
||||
|
||||
// keep following chain
|
||||
if targetResult.rcode == -1 {
|
||||
cnameQuestion = dns.Question{Name: targetName, Qtype: dns.TypeCNAME, Qclass: cnameQuestion.Qclass}
|
||||
continue
|
||||
}
|
||||
|
||||
return d.buildChainResult(chain, targetResult)
|
||||
}
|
||||
|
||||
if len(chain) > 0 {
|
||||
return lookupResult{records: chain, rcode: dns.RcodeSuccess}
|
||||
}
|
||||
return lookupResult{rcode: dns.RcodeSuccess}
|
||||
}
|
||||
|
||||
// buildChainResult combines CNAME chain records with the target resolution result.
|
||||
// Per RFC 6604, the final rcode is propagated through the chain.
|
||||
func (d *Resolver) buildChainResult(chain []dns.RR, target lookupResult) lookupResult {
|
||||
records := chain
|
||||
if len(target.records) > 0 {
|
||||
records = append(records, target.records...)
|
||||
}
|
||||
|
||||
// preserve hasExternalData for SERVFAIL so caller knows the error came from upstream
|
||||
if target.hasExternalData && target.rcode == dns.RcodeServerFailure {
|
||||
return lookupResult{
|
||||
records: records,
|
||||
rcode: dns.RcodeServerFailure,
|
||||
hasExternalData: true,
|
||||
}
|
||||
}
|
||||
|
||||
return lookupResult{
|
||||
records: records,
|
||||
rcode: target.rcode,
|
||||
hasExternalData: target.hasExternalData,
|
||||
}
|
||||
}
|
||||
|
||||
// resolveCNAMETarget attempts to resolve a CNAME target name.
|
||||
// Returns rcode=-1 to signal "keep following the chain".
|
||||
func (d *Resolver) resolveCNAMETarget(logger *log.Entry, targetName string, targetType uint16, qclass uint16) lookupResult {
|
||||
if records := d.getRecords(dns.Question{Name: targetName, Qtype: targetType, Qclass: qclass}); len(records) > 0 {
|
||||
return lookupResult{records: records, rcode: dns.RcodeSuccess}
|
||||
}
|
||||
|
||||
// another CNAME, keep following
|
||||
if d.hasRecord(dns.Question{Name: targetName, Qtype: dns.TypeCNAME, Qclass: qclass}) {
|
||||
return lookupResult{rcode: -1}
|
||||
}
|
||||
|
||||
// domain exists locally but not this record type (NODATA)
|
||||
if d.hasRecordsForDomain(domain.Domain(targetName)) {
|
||||
return lookupResult{rcode: dns.RcodeSuccess}
|
||||
}
|
||||
|
||||
// in our zone but doesn't exist (NXDOMAIN)
|
||||
if d.isInManagedZone(targetName) {
|
||||
return lookupResult{rcode: dns.RcodeNameError}
|
||||
}
|
||||
|
||||
return d.resolveExternal(logger, targetName, targetType)
|
||||
}
|
||||
|
||||
func (d *Resolver) getRecords(q dns.Question) []dns.RR {
|
||||
d.mu.RLock()
|
||||
defer d.mu.RUnlock()
|
||||
return d.records[q]
|
||||
}
|
||||
|
||||
func (d *Resolver) hasRecord(q dns.Question) bool {
|
||||
d.mu.RLock()
|
||||
defer d.mu.RUnlock()
|
||||
_, ok := d.records[q]
|
||||
return ok
|
||||
}
|
||||
|
||||
// resolveExternal resolves a domain name using the system resolver.
|
||||
// This is used to resolve CNAME targets that point outside our local zone,
|
||||
// which is required for musl libc compatibility (musl expects complete answers).
|
||||
func (d *Resolver) resolveExternal(logger *log.Entry, name string, qtype uint16) lookupResult {
|
||||
network := resutil.NetworkForQtype(qtype)
|
||||
if network == "" {
|
||||
return lookupResult{rcode: dns.RcodeNotImplemented}
|
||||
}
|
||||
|
||||
resolver := d.resolver
|
||||
if resolver == nil {
|
||||
resolver = net.DefaultResolver
|
||||
}
|
||||
|
||||
ctx, cancel := context.WithTimeout(d.ctx, externalResolutionTimeout)
|
||||
defer cancel()
|
||||
|
||||
result := resutil.LookupIP(ctx, resolver, network, name, qtype)
|
||||
if result.Err != nil {
|
||||
d.logDNSError(logger, name, qtype, result.Err)
|
||||
return lookupResult{rcode: result.Rcode, hasExternalData: true}
|
||||
}
|
||||
|
||||
return lookupResult{
|
||||
records: resutil.IPsToRRs(name, result.IPs, 60),
|
||||
rcode: dns.RcodeSuccess,
|
||||
hasExternalData: true,
|
||||
}
|
||||
}
|
||||
|
||||
// logDNSError logs DNS resolution errors for debugging.
|
||||
func (d *Resolver) logDNSError(logger *log.Entry, hostname string, qtype uint16, err error) {
|
||||
qtypeName := dns.TypeToString[qtype]
|
||||
|
||||
var dnsErr *net.DNSError
|
||||
if !errors.As(err, &dnsErr) {
|
||||
logger.Debugf("DNS resolution failed for %s type %s: %v", hostname, qtypeName, err)
|
||||
return
|
||||
}
|
||||
|
||||
if dnsErr.IsNotFound {
|
||||
logger.Tracef("DNS target not found: %s type %s", hostname, qtypeName)
|
||||
return
|
||||
}
|
||||
|
||||
if dnsErr.Server != "" {
|
||||
logger.Debugf("DNS resolution failed for %s type %s server=%s: %v", hostname, qtypeName, dnsErr.Server, err)
|
||||
} else {
|
||||
logger.Debugf("DNS resolution failed for %s type %s: %v", hostname, qtypeName, err)
|
||||
}
|
||||
}
|
||||
|
||||
// Update replaces all zones and their records
|
||||
func (d *Resolver) Update(customZones []nbdns.CustomZone) {
|
||||
func (d *Resolver) Update(update []nbdns.SimpleRecord) {
|
||||
d.mu.Lock()
|
||||
defer d.mu.Unlock()
|
||||
|
||||
maps.Clear(d.records)
|
||||
maps.Clear(d.domains)
|
||||
maps.Clear(d.zones)
|
||||
|
||||
for _, zone := range customZones {
|
||||
zoneDomain := domain.Domain(strings.ToLower(dns.Fqdn(zone.Domain)))
|
||||
d.zones[zoneDomain] = zone.NonAuthoritative
|
||||
|
||||
for _, rec := range zone.Records {
|
||||
if err := d.registerRecord(rec); err != nil {
|
||||
log.Warnf("failed to register the record (%s): %v", rec, err)
|
||||
}
|
||||
for _, rec := range update {
|
||||
if err := d.registerRecord(rec); err != nil {
|
||||
log.Warnf("failed to register the record (%s): %v", rec, err)
|
||||
continue
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
@@ -1,14 +1,8 @@
|
||||
package local
|
||||
|
||||
import (
|
||||
"context"
|
||||
"fmt"
|
||||
"net"
|
||||
"net/netip"
|
||||
"strings"
|
||||
"sync"
|
||||
"testing"
|
||||
"time"
|
||||
|
||||
"github.com/miekg/dns"
|
||||
"github.com/stretchr/testify/assert"
|
||||
@@ -18,18 +12,6 @@ import (
|
||||
nbdns "github.com/netbirdio/netbird/dns"
|
||||
)
|
||||
|
||||
// mockResolver implements resolver for testing
|
||||
type mockResolver struct {
|
||||
lookupFunc func(ctx context.Context, network, host string) ([]netip.Addr, error)
|
||||
}
|
||||
|
||||
func (m *mockResolver) LookupNetIP(ctx context.Context, network, host string) ([]netip.Addr, error) {
|
||||
if m.lookupFunc != nil {
|
||||
return m.lookupFunc(ctx, network, host)
|
||||
}
|
||||
return nil, nil
|
||||
}
|
||||
|
||||
func TestLocalResolver_ServeDNS(t *testing.T) {
|
||||
recordA := nbdns.SimpleRecord{
|
||||
Name: "peera.netbird.cloud.",
|
||||
@@ -124,11 +106,11 @@ func TestLocalResolver_Update_StaleRecord(t *testing.T) {
|
||||
|
||||
resolver := NewResolver()
|
||||
|
||||
zone1 := []nbdns.CustomZone{{Domain: "example.com.", Records: []nbdns.SimpleRecord{record1}}}
|
||||
zone2 := []nbdns.CustomZone{{Domain: "example.com.", Records: []nbdns.SimpleRecord{record2}}}
|
||||
update1 := []nbdns.SimpleRecord{record1}
|
||||
update2 := []nbdns.SimpleRecord{record2}
|
||||
|
||||
// Apply first update
|
||||
resolver.Update(zone1)
|
||||
resolver.Update(update1)
|
||||
|
||||
// Verify first update
|
||||
resolver.mu.RLock()
|
||||
@@ -140,7 +122,7 @@ func TestLocalResolver_Update_StaleRecord(t *testing.T) {
|
||||
assert.Contains(t, rrSlice1[0].String(), record1.RData, "Record after first update should be %s", record1.RData)
|
||||
|
||||
// Apply second update
|
||||
resolver.Update(zone2)
|
||||
resolver.Update(update2)
|
||||
|
||||
// Verify second update
|
||||
resolver.mu.RLock()
|
||||
@@ -169,10 +151,10 @@ func TestLocalResolver_MultipleRecords_SameQuestion(t *testing.T) {
|
||||
Name: recordName, Type: int(recordType), Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.2",
|
||||
}
|
||||
|
||||
zones := []nbdns.CustomZone{{Domain: "example.com.", Records: []nbdns.SimpleRecord{record1, record2}}}
|
||||
update := []nbdns.SimpleRecord{record1, record2}
|
||||
|
||||
// Apply update with both records
|
||||
resolver.Update(zones)
|
||||
resolver.Update(update)
|
||||
|
||||
// Create question that matches both records
|
||||
question := dns.Question{
|
||||
@@ -213,10 +195,10 @@ func TestLocalResolver_RecordRotation(t *testing.T) {
|
||||
Name: recordName, Type: int(recordType), Class: nbdns.DefaultClass, TTL: 300, RData: "192.168.1.3",
|
||||
}
|
||||
|
||||
zones := []nbdns.CustomZone{{Domain: "example.com.", Records: []nbdns.SimpleRecord{record1, record2, record3}}}
|
||||
update := []nbdns.SimpleRecord{record1, record2, record3}
|
||||
|
||||
// Apply update with all three records
|
||||
resolver.Update(zones)
|
||||
resolver.Update(update)
|
||||
|
||||
msg := new(dns.Msg).SetQuestion(recordName, recordType)
|
||||
|
||||
@@ -282,7 +264,7 @@ func TestLocalResolver_CaseInsensitiveMatching(t *testing.T) {
|
||||
}
|
||||
|
||||
// Update resolver with the records
|
||||
resolver.Update([]nbdns.CustomZone{{Domain: "example.com.", Records: []nbdns.SimpleRecord{lowerCaseRecord, mixedCaseRecord}}})
|
||||
resolver.Update([]nbdns.SimpleRecord{lowerCaseRecord, mixedCaseRecord})
|
||||
|
||||
testCases := []struct {
|
||||
name string
|
||||
@@ -397,7 +379,7 @@ func TestLocalResolver_CNAMEFallback(t *testing.T) {
|
||||
}
|
||||
|
||||
// Update resolver with both records
|
||||
resolver.Update([]nbdns.CustomZone{{Domain: "example.com.", Records: []nbdns.SimpleRecord{cnameRecord, targetRecord}}})
|
||||
resolver.Update([]nbdns.SimpleRecord{cnameRecord, targetRecord})
|
||||
|
||||
testCases := []struct {
|
||||
name string
|
||||
@@ -494,20 +476,6 @@ func TestLocalResolver_CNAMEFallback(t *testing.T) {
|
||||
// with 0 records instead of NXDOMAIN
|
||||
func TestLocalResolver_NoErrorWithDifferentRecordType(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
// Mock external resolver for CNAME target resolution
|
||||
resolver.resolver = &mockResolver{
|
||||
lookupFunc: func(_ context.Context, network, host string) ([]netip.Addr, error) {
|
||||
if host == "target.example.com." {
|
||||
if network == "ip4" {
|
||||
return []netip.Addr{netip.MustParseAddr("93.184.216.34")}, nil
|
||||
}
|
||||
if network == "ip6" {
|
||||
return []netip.Addr{netip.MustParseAddr("2606:2800:220:1:248:1893:25c8:1946")}, nil
|
||||
}
|
||||
}
|
||||
return nil, &net.DNSError{IsNotFound: true, Name: host}
|
||||
},
|
||||
}
|
||||
|
||||
recordA := nbdns.SimpleRecord{
|
||||
Name: "example.netbird.cloud.",
|
||||
@@ -525,7 +493,7 @@ func TestLocalResolver_NoErrorWithDifferentRecordType(t *testing.T) {
|
||||
RData: "target.example.com.",
|
||||
}
|
||||
|
||||
resolver.Update([]nbdns.CustomZone{{Domain: "netbird.cloud.", Records: []nbdns.SimpleRecord{recordA, recordCNAME}}})
|
||||
resolver.Update([]nbdns.SimpleRecord{recordA, recordCNAME})
|
||||
|
||||
testCases := []struct {
|
||||
name string
|
||||
@@ -614,808 +582,3 @@ func TestLocalResolver_NoErrorWithDifferentRecordType(t *testing.T) {
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
// TestLocalResolver_CNAMEChainResolution tests comprehensive CNAME chain following
|
||||
func TestLocalResolver_CNAMEChainResolution(t *testing.T) {
|
||||
t.Run("simple internal CNAME chain", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
resolver.Update([]nbdns.CustomZone{{
|
||||
Domain: "example.com.",
|
||||
Records: []nbdns.SimpleRecord{
|
||||
{Name: "alias.example.com.", Type: int(dns.TypeCNAME), Class: nbdns.DefaultClass, TTL: 300, RData: "target.example.com."},
|
||||
{Name: "target.example.com.", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "192.168.1.1"},
|
||||
},
|
||||
}})
|
||||
|
||||
msg := new(dns.Msg).SetQuestion("alias.example.com.", dns.TypeA)
|
||||
var resp *dns.Msg
|
||||
resolver.ServeDNS(&test.MockResponseWriter{WriteMsgFunc: func(m *dns.Msg) error { resp = m; return nil }}, msg)
|
||||
|
||||
require.NotNil(t, resp)
|
||||
require.Equal(t, dns.RcodeSuccess, resp.Rcode)
|
||||
require.Len(t, resp.Answer, 2)
|
||||
|
||||
cname, ok := resp.Answer[0].(*dns.CNAME)
|
||||
require.True(t, ok)
|
||||
assert.Equal(t, "target.example.com.", cname.Target)
|
||||
|
||||
a, ok := resp.Answer[1].(*dns.A)
|
||||
require.True(t, ok)
|
||||
assert.Equal(t, "192.168.1.1", a.A.String())
|
||||
})
|
||||
|
||||
t.Run("multi-hop CNAME chain", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
resolver.Update([]nbdns.CustomZone{{
|
||||
Domain: "test.",
|
||||
Records: []nbdns.SimpleRecord{
|
||||
{Name: "hop1.test.", Type: int(dns.TypeCNAME), Class: nbdns.DefaultClass, TTL: 300, RData: "hop2.test."},
|
||||
{Name: "hop2.test.", Type: int(dns.TypeCNAME), Class: nbdns.DefaultClass, TTL: 300, RData: "hop3.test."},
|
||||
{Name: "hop3.test.", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"},
|
||||
},
|
||||
}})
|
||||
|
||||
msg := new(dns.Msg).SetQuestion("hop1.test.", dns.TypeA)
|
||||
var resp *dns.Msg
|
||||
resolver.ServeDNS(&test.MockResponseWriter{WriteMsgFunc: func(m *dns.Msg) error { resp = m; return nil }}, msg)
|
||||
|
||||
require.NotNil(t, resp)
|
||||
require.Equal(t, dns.RcodeSuccess, resp.Rcode)
|
||||
require.Len(t, resp.Answer, 3)
|
||||
})
|
||||
|
||||
t.Run("CNAME to non-existent internal target returns only CNAME", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
resolver.Update([]nbdns.CustomZone{{
|
||||
Domain: "test.",
|
||||
Records: []nbdns.SimpleRecord{
|
||||
{Name: "alias.test.", Type: int(dns.TypeCNAME), Class: nbdns.DefaultClass, TTL: 300, RData: "nonexistent.test."},
|
||||
},
|
||||
}})
|
||||
|
||||
msg := new(dns.Msg).SetQuestion("alias.test.", dns.TypeA)
|
||||
var resp *dns.Msg
|
||||
resolver.ServeDNS(&test.MockResponseWriter{WriteMsgFunc: func(m *dns.Msg) error { resp = m; return nil }}, msg)
|
||||
|
||||
require.NotNil(t, resp)
|
||||
require.Len(t, resp.Answer, 1)
|
||||
_, ok := resp.Answer[0].(*dns.CNAME)
|
||||
assert.True(t, ok)
|
||||
})
|
||||
}
|
||||
|
||||
// TestLocalResolver_CNAMEMaxDepth tests the maximum depth limit for CNAME chains
|
||||
func TestLocalResolver_CNAMEMaxDepth(t *testing.T) {
|
||||
t.Run("chain at max depth resolves", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
var records []nbdns.SimpleRecord
|
||||
// Create chain of 7 CNAMEs (under max of 8)
|
||||
for i := 1; i <= 7; i++ {
|
||||
records = append(records, nbdns.SimpleRecord{
|
||||
Name: fmt.Sprintf("hop%d.test.", i),
|
||||
Type: int(dns.TypeCNAME),
|
||||
Class: nbdns.DefaultClass,
|
||||
TTL: 300,
|
||||
RData: fmt.Sprintf("hop%d.test.", i+1),
|
||||
})
|
||||
}
|
||||
records = append(records, nbdns.SimpleRecord{
|
||||
Name: "hop8.test.", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "10.10.10.10",
|
||||
})
|
||||
|
||||
resolver.Update([]nbdns.CustomZone{{Domain: "test.", Records: records}})
|
||||
|
||||
msg := new(dns.Msg).SetQuestion("hop1.test.", dns.TypeA)
|
||||
var resp *dns.Msg
|
||||
resolver.ServeDNS(&test.MockResponseWriter{WriteMsgFunc: func(m *dns.Msg) error { resp = m; return nil }}, msg)
|
||||
|
||||
require.NotNil(t, resp)
|
||||
require.Equal(t, dns.RcodeSuccess, resp.Rcode)
|
||||
require.Len(t, resp.Answer, 8)
|
||||
})
|
||||
|
||||
t.Run("chain exceeding max depth stops", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
var records []nbdns.SimpleRecord
|
||||
// Create chain of 10 CNAMEs (exceeds max of 8)
|
||||
for i := 1; i <= 10; i++ {
|
||||
records = append(records, nbdns.SimpleRecord{
|
||||
Name: fmt.Sprintf("deep%d.test.", i),
|
||||
Type: int(dns.TypeCNAME),
|
||||
Class: nbdns.DefaultClass,
|
||||
TTL: 300,
|
||||
RData: fmt.Sprintf("deep%d.test.", i+1),
|
||||
})
|
||||
}
|
||||
records = append(records, nbdns.SimpleRecord{
|
||||
Name: "deep11.test.", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "10.10.10.10",
|
||||
})
|
||||
|
||||
resolver.Update([]nbdns.CustomZone{{Domain: "test.", Records: records}})
|
||||
|
||||
msg := new(dns.Msg).SetQuestion("deep1.test.", dns.TypeA)
|
||||
var resp *dns.Msg
|
||||
resolver.ServeDNS(&test.MockResponseWriter{WriteMsgFunc: func(m *dns.Msg) error { resp = m; return nil }}, msg)
|
||||
|
||||
require.NotNil(t, resp)
|
||||
// Should NOT have the final A record (chain too deep)
|
||||
assert.LessOrEqual(t, len(resp.Answer), 8)
|
||||
})
|
||||
|
||||
t.Run("circular CNAME is protected by max depth", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
resolver.Update([]nbdns.CustomZone{{
|
||||
Domain: "test.",
|
||||
Records: []nbdns.SimpleRecord{
|
||||
{Name: "loop1.test.", Type: int(dns.TypeCNAME), Class: nbdns.DefaultClass, TTL: 300, RData: "loop2.test."},
|
||||
{Name: "loop2.test.", Type: int(dns.TypeCNAME), Class: nbdns.DefaultClass, TTL: 300, RData: "loop1.test."},
|
||||
},
|
||||
}})
|
||||
|
||||
msg := new(dns.Msg).SetQuestion("loop1.test.", dns.TypeA)
|
||||
var resp *dns.Msg
|
||||
resolver.ServeDNS(&test.MockResponseWriter{WriteMsgFunc: func(m *dns.Msg) error { resp = m; return nil }}, msg)
|
||||
|
||||
require.NotNil(t, resp)
|
||||
assert.LessOrEqual(t, len(resp.Answer), 8)
|
||||
})
|
||||
}
|
||||
|
||||
// TestLocalResolver_ExternalCNAMEResolution tests CNAME resolution to external domains
|
||||
func TestLocalResolver_ExternalCNAMEResolution(t *testing.T) {
|
||||
t.Run("CNAME to external domain resolves via external resolver", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
resolver.resolver = &mockResolver{
|
||||
lookupFunc: func(_ context.Context, network, host string) ([]netip.Addr, error) {
|
||||
if host == "external.example.com." && network == "ip4" {
|
||||
return []netip.Addr{netip.MustParseAddr("93.184.216.34")}, nil
|
||||
}
|
||||
return nil, nil
|
||||
},
|
||||
}
|
||||
|
||||
resolver.Update([]nbdns.CustomZone{{
|
||||
Domain: "test.",
|
||||
Records: []nbdns.SimpleRecord{
|
||||
{Name: "alias.test.", Type: int(dns.TypeCNAME), Class: nbdns.DefaultClass, TTL: 300, RData: "external.example.com."},
|
||||
},
|
||||
}})
|
||||
|
||||
msg := new(dns.Msg).SetQuestion("alias.test.", dns.TypeA)
|
||||
var resp *dns.Msg
|
||||
resolver.ServeDNS(&test.MockResponseWriter{WriteMsgFunc: func(m *dns.Msg) error { resp = m; return nil }}, msg)
|
||||
|
||||
require.NotNil(t, resp)
|
||||
require.Len(t, resp.Answer, 2, "Should have CNAME + A record")
|
||||
|
||||
cname, ok := resp.Answer[0].(*dns.CNAME)
|
||||
require.True(t, ok)
|
||||
assert.Equal(t, "external.example.com.", cname.Target)
|
||||
|
||||
a, ok := resp.Answer[1].(*dns.A)
|
||||
require.True(t, ok)
|
||||
assert.Equal(t, "93.184.216.34", a.A.String())
|
||||
})
|
||||
|
||||
t.Run("CNAME to external domain resolves IPv6", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
resolver.resolver = &mockResolver{
|
||||
lookupFunc: func(_ context.Context, network, host string) ([]netip.Addr, error) {
|
||||
if host == "external.example.com." && network == "ip6" {
|
||||
return []netip.Addr{netip.MustParseAddr("2606:2800:220:1:248:1893:25c8:1946")}, nil
|
||||
}
|
||||
return nil, nil
|
||||
},
|
||||
}
|
||||
|
||||
resolver.Update([]nbdns.CustomZone{{
|
||||
Domain: "test.",
|
||||
Records: []nbdns.SimpleRecord{
|
||||
{Name: "alias.test.", Type: int(dns.TypeCNAME), Class: nbdns.DefaultClass, TTL: 300, RData: "external.example.com."},
|
||||
},
|
||||
}})
|
||||
|
||||
msg := new(dns.Msg).SetQuestion("alias.test.", dns.TypeAAAA)
|
||||
var resp *dns.Msg
|
||||
resolver.ServeDNS(&test.MockResponseWriter{WriteMsgFunc: func(m *dns.Msg) error { resp = m; return nil }}, msg)
|
||||
|
||||
require.NotNil(t, resp)
|
||||
require.Len(t, resp.Answer, 2, "Should have CNAME + AAAA record")
|
||||
|
||||
cname, ok := resp.Answer[0].(*dns.CNAME)
|
||||
require.True(t, ok)
|
||||
assert.Equal(t, "external.example.com.", cname.Target)
|
||||
|
||||
aaaa, ok := resp.Answer[1].(*dns.AAAA)
|
||||
require.True(t, ok)
|
||||
assert.Equal(t, "2606:2800:220:1:248:1893:25c8:1946", aaaa.AAAA.String())
|
||||
})
|
||||
|
||||
t.Run("concurrent external resolution", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
resolver.resolver = &mockResolver{
|
||||
lookupFunc: func(_ context.Context, network, host string) ([]netip.Addr, error) {
|
||||
if host == "external.example.com." && network == "ip4" {
|
||||
return []netip.Addr{netip.MustParseAddr("93.184.216.34")}, nil
|
||||
}
|
||||
return nil, nil
|
||||
},
|
||||
}
|
||||
|
||||
resolver.Update([]nbdns.CustomZone{{
|
||||
Domain: "test.",
|
||||
Records: []nbdns.SimpleRecord{
|
||||
{Name: "concurrent.test.", Type: int(dns.TypeCNAME), Class: nbdns.DefaultClass, TTL: 300, RData: "external.example.com."},
|
||||
},
|
||||
}})
|
||||
|
||||
var wg sync.WaitGroup
|
||||
results := make([]*dns.Msg, 10)
|
||||
|
||||
for i := 0; i < 10; i++ {
|
||||
wg.Add(1)
|
||||
go func(idx int) {
|
||||
defer wg.Done()
|
||||
msg := new(dns.Msg).SetQuestion("concurrent.test.", dns.TypeA)
|
||||
var resp *dns.Msg
|
||||
resolver.ServeDNS(&test.MockResponseWriter{WriteMsgFunc: func(m *dns.Msg) error { resp = m; return nil }}, msg)
|
||||
results[idx] = resp
|
||||
}(i)
|
||||
}
|
||||
wg.Wait()
|
||||
|
||||
for i, resp := range results {
|
||||
require.NotNil(t, resp, "Response %d should not be nil", i)
|
||||
require.Len(t, resp.Answer, 2, "Response %d should have CNAME + A", i)
|
||||
}
|
||||
})
|
||||
}
|
||||
|
||||
// TestLocalResolver_ZoneManagement tests zone-aware CNAME resolution
|
||||
func TestLocalResolver_ZoneManagement(t *testing.T) {
|
||||
t.Run("Update sets zones correctly", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
|
||||
resolver.Update([]nbdns.CustomZone{
|
||||
{Domain: "example.com.", Records: []nbdns.SimpleRecord{
|
||||
{Name: "host.example.com.", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"},
|
||||
}},
|
||||
{Domain: "test.local."},
|
||||
})
|
||||
|
||||
assert.True(t, resolver.isInManagedZone("host.example.com."))
|
||||
assert.True(t, resolver.isInManagedZone("other.example.com."))
|
||||
assert.True(t, resolver.isInManagedZone("sub.test.local."))
|
||||
assert.False(t, resolver.isInManagedZone("external.com."))
|
||||
})
|
||||
|
||||
t.Run("isInManagedZone case insensitive", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
resolver.Update([]nbdns.CustomZone{{Domain: "Example.COM."}})
|
||||
|
||||
assert.True(t, resolver.isInManagedZone("host.example.com."))
|
||||
assert.True(t, resolver.isInManagedZone("HOST.EXAMPLE.COM."))
|
||||
})
|
||||
|
||||
t.Run("Update clears zones", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
resolver.Update([]nbdns.CustomZone{{Domain: "example.com."}})
|
||||
assert.True(t, resolver.isInManagedZone("host.example.com."))
|
||||
|
||||
resolver.Update(nil)
|
||||
assert.False(t, resolver.isInManagedZone("host.example.com."))
|
||||
})
|
||||
}
|
||||
|
||||
// TestLocalResolver_CNAMEZoneAwareResolution tests CNAME resolution with zone awareness
|
||||
func TestLocalResolver_CNAMEZoneAwareResolution(t *testing.T) {
|
||||
t.Run("CNAME target in managed zone returns NXDOMAIN per RFC 6604", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
resolver.Update([]nbdns.CustomZone{{
|
||||
Domain: "myzone.test.",
|
||||
Records: []nbdns.SimpleRecord{
|
||||
{Name: "alias.myzone.test.", Type: int(dns.TypeCNAME), Class: nbdns.DefaultClass, TTL: 300, RData: "nonexistent.myzone.test."},
|
||||
},
|
||||
}})
|
||||
|
||||
msg := new(dns.Msg).SetQuestion("alias.myzone.test.", dns.TypeA)
|
||||
var resp *dns.Msg
|
||||
resolver.ServeDNS(&test.MockResponseWriter{WriteMsgFunc: func(m *dns.Msg) error { resp = m; return nil }}, msg)
|
||||
|
||||
require.NotNil(t, resp)
|
||||
assert.Equal(t, dns.RcodeNameError, resp.Rcode, "Should return NXDOMAIN")
|
||||
require.Len(t, resp.Answer, 1, "Should include CNAME in answer")
|
||||
})
|
||||
|
||||
t.Run("CNAME to external domain skips zone check", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
resolver.resolver = &mockResolver{
|
||||
lookupFunc: func(_ context.Context, network, host string) ([]netip.Addr, error) {
|
||||
if host == "external.other.com." && network == "ip4" {
|
||||
return []netip.Addr{netip.MustParseAddr("203.0.113.1")}, nil
|
||||
}
|
||||
return nil, nil
|
||||
},
|
||||
}
|
||||
|
||||
resolver.Update([]nbdns.CustomZone{{
|
||||
Domain: "myzone.test.",
|
||||
Records: []nbdns.SimpleRecord{
|
||||
{Name: "alias.myzone.test.", Type: int(dns.TypeCNAME), Class: nbdns.DefaultClass, TTL: 300, RData: "external.other.com."},
|
||||
},
|
||||
}})
|
||||
|
||||
msg := new(dns.Msg).SetQuestion("alias.myzone.test.", dns.TypeA)
|
||||
var resp *dns.Msg
|
||||
resolver.ServeDNS(&test.MockResponseWriter{WriteMsgFunc: func(m *dns.Msg) error { resp = m; return nil }}, msg)
|
||||
|
||||
require.NotNil(t, resp)
|
||||
assert.Equal(t, dns.RcodeSuccess, resp.Rcode)
|
||||
require.Len(t, resp.Answer, 2, "Should have CNAME + A from external resolution")
|
||||
})
|
||||
|
||||
t.Run("CNAME target exists with different type returns NODATA not NXDOMAIN", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
// CNAME points to target that has A but no AAAA - query for AAAA should be NODATA
|
||||
resolver.Update([]nbdns.CustomZone{{
|
||||
Domain: "myzone.test.",
|
||||
Records: []nbdns.SimpleRecord{
|
||||
{Name: "alias.myzone.test.", Type: int(dns.TypeCNAME), Class: nbdns.DefaultClass, TTL: 300, RData: "target.myzone.test."},
|
||||
{Name: "target.myzone.test.", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "1.1.1.1"},
|
||||
},
|
||||
}})
|
||||
|
||||
msg := new(dns.Msg).SetQuestion("alias.myzone.test.", dns.TypeAAAA)
|
||||
var resp *dns.Msg
|
||||
resolver.ServeDNS(&test.MockResponseWriter{WriteMsgFunc: func(m *dns.Msg) error { resp = m; return nil }}, msg)
|
||||
|
||||
require.NotNil(t, resp)
|
||||
assert.Equal(t, dns.RcodeSuccess, resp.Rcode, "Should return NODATA (success), not NXDOMAIN")
|
||||
require.Len(t, resp.Answer, 1, "Should have only CNAME, no AAAA")
|
||||
_, ok := resp.Answer[0].(*dns.CNAME)
|
||||
assert.True(t, ok, "Answer should be CNAME record")
|
||||
})
|
||||
|
||||
t.Run("external CNAME target exists but no AAAA records (NODATA)", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
resolver.resolver = &mockResolver{
|
||||
lookupFunc: func(_ context.Context, network, host string) ([]netip.Addr, error) {
|
||||
if host == "external.example.com." {
|
||||
if network == "ip6" {
|
||||
// No AAAA records
|
||||
return nil, &net.DNSError{IsNotFound: true, Name: host}
|
||||
}
|
||||
if network == "ip4" {
|
||||
// But A records exist - domain exists
|
||||
return []netip.Addr{netip.MustParseAddr("93.184.216.34")}, nil
|
||||
}
|
||||
}
|
||||
return nil, &net.DNSError{IsNotFound: true, Name: host}
|
||||
},
|
||||
}
|
||||
|
||||
resolver.Update([]nbdns.CustomZone{{
|
||||
Domain: "test.",
|
||||
Records: []nbdns.SimpleRecord{
|
||||
{Name: "alias.test.", Type: int(dns.TypeCNAME), Class: nbdns.DefaultClass, TTL: 300, RData: "external.example.com."},
|
||||
},
|
||||
}})
|
||||
|
||||
msg := new(dns.Msg).SetQuestion("alias.test.", dns.TypeAAAA)
|
||||
var resp *dns.Msg
|
||||
resolver.ServeDNS(&test.MockResponseWriter{WriteMsgFunc: func(m *dns.Msg) error { resp = m; return nil }}, msg)
|
||||
|
||||
require.NotNil(t, resp)
|
||||
assert.Equal(t, dns.RcodeSuccess, resp.Rcode, "Should return NODATA (success), not NXDOMAIN")
|
||||
require.Len(t, resp.Answer, 1, "Should have only CNAME")
|
||||
_, ok := resp.Answer[0].(*dns.CNAME)
|
||||
assert.True(t, ok, "Answer should be CNAME record")
|
||||
})
|
||||
|
||||
// Table-driven test for all external resolution outcomes
|
||||
externalCases := []struct {
|
||||
name string
|
||||
lookupFunc func(context.Context, string, string) ([]netip.Addr, error)
|
||||
expectedRcode int
|
||||
expectedAnswer int
|
||||
}{
|
||||
{
|
||||
name: "external NXDOMAIN (both A and AAAA not found)",
|
||||
lookupFunc: func(_ context.Context, network, host string) ([]netip.Addr, error) {
|
||||
return nil, &net.DNSError{IsNotFound: true, Name: host}
|
||||
},
|
||||
expectedRcode: dns.RcodeNameError,
|
||||
expectedAnswer: 1, // CNAME only
|
||||
},
|
||||
{
|
||||
name: "external SERVFAIL (temporary error)",
|
||||
lookupFunc: func(_ context.Context, network, host string) ([]netip.Addr, error) {
|
||||
return nil, &net.DNSError{IsTemporary: true, Name: host}
|
||||
},
|
||||
expectedRcode: dns.RcodeServerFailure,
|
||||
expectedAnswer: 1, // CNAME only
|
||||
},
|
||||
{
|
||||
name: "external SERVFAIL (timeout)",
|
||||
lookupFunc: func(_ context.Context, network, host string) ([]netip.Addr, error) {
|
||||
return nil, &net.DNSError{IsTimeout: true, Name: host}
|
||||
},
|
||||
expectedRcode: dns.RcodeServerFailure,
|
||||
expectedAnswer: 1, // CNAME only
|
||||
},
|
||||
{
|
||||
name: "external SERVFAIL (generic error)",
|
||||
lookupFunc: func(_ context.Context, network, host string) ([]netip.Addr, error) {
|
||||
return nil, fmt.Errorf("connection refused")
|
||||
},
|
||||
expectedRcode: dns.RcodeServerFailure,
|
||||
expectedAnswer: 1, // CNAME only
|
||||
},
|
||||
{
|
||||
name: "external success with IPs",
|
||||
lookupFunc: func(_ context.Context, network, host string) ([]netip.Addr, error) {
|
||||
if network == "ip4" {
|
||||
return []netip.Addr{netip.MustParseAddr("93.184.216.34")}, nil
|
||||
}
|
||||
return nil, &net.DNSError{IsNotFound: true, Name: host}
|
||||
},
|
||||
expectedRcode: dns.RcodeSuccess,
|
||||
expectedAnswer: 2, // CNAME + A
|
||||
},
|
||||
}
|
||||
|
||||
for _, tc := range externalCases {
|
||||
t.Run(tc.name, func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
resolver.resolver = &mockResolver{lookupFunc: tc.lookupFunc}
|
||||
|
||||
resolver.Update([]nbdns.CustomZone{{
|
||||
Domain: "test.",
|
||||
Records: []nbdns.SimpleRecord{
|
||||
{Name: "alias.test.", Type: int(dns.TypeCNAME), Class: nbdns.DefaultClass, TTL: 300, RData: "external.example.com."},
|
||||
},
|
||||
}})
|
||||
|
||||
msg := new(dns.Msg).SetQuestion("alias.test.", dns.TypeA)
|
||||
var resp *dns.Msg
|
||||
resolver.ServeDNS(&test.MockResponseWriter{WriteMsgFunc: func(m *dns.Msg) error { resp = m; return nil }}, msg)
|
||||
|
||||
require.NotNil(t, resp)
|
||||
assert.Equal(t, tc.expectedRcode, resp.Rcode, "rcode mismatch")
|
||||
assert.Len(t, resp.Answer, tc.expectedAnswer, "answer count mismatch")
|
||||
if tc.expectedAnswer > 0 {
|
||||
_, ok := resp.Answer[0].(*dns.CNAME)
|
||||
assert.True(t, ok, "first answer should be CNAME")
|
||||
}
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
// TestLocalResolver_Fallthrough verifies that non-authoritative zones
|
||||
// trigger fallthrough (Zero bit set) when no records match
|
||||
func TestLocalResolver_Fallthrough(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
|
||||
record := nbdns.SimpleRecord{
|
||||
Name: "existing.custom.zone.",
|
||||
Type: int(dns.TypeA),
|
||||
Class: nbdns.DefaultClass,
|
||||
TTL: 300,
|
||||
RData: "10.0.0.1",
|
||||
}
|
||||
|
||||
testCases := []struct {
|
||||
name string
|
||||
zones []nbdns.CustomZone
|
||||
queryName string
|
||||
expectFallthrough bool
|
||||
expectRecord bool
|
||||
}{
|
||||
{
|
||||
name: "Authoritative zone returns NXDOMAIN without fallthrough",
|
||||
zones: []nbdns.CustomZone{{
|
||||
Domain: "custom.zone.",
|
||||
Records: []nbdns.SimpleRecord{record},
|
||||
}},
|
||||
queryName: "nonexistent.custom.zone.",
|
||||
expectFallthrough: false,
|
||||
expectRecord: false,
|
||||
},
|
||||
{
|
||||
name: "Non-authoritative zone triggers fallthrough",
|
||||
zones: []nbdns.CustomZone{{
|
||||
Domain: "custom.zone.",
|
||||
Records: []nbdns.SimpleRecord{record},
|
||||
NonAuthoritative: true,
|
||||
}},
|
||||
queryName: "nonexistent.custom.zone.",
|
||||
expectFallthrough: true,
|
||||
expectRecord: false,
|
||||
},
|
||||
{
|
||||
name: "Record found in non-authoritative zone returns normally",
|
||||
zones: []nbdns.CustomZone{{
|
||||
Domain: "custom.zone.",
|
||||
Records: []nbdns.SimpleRecord{record},
|
||||
NonAuthoritative: true,
|
||||
}},
|
||||
queryName: "existing.custom.zone.",
|
||||
expectFallthrough: false,
|
||||
expectRecord: true,
|
||||
},
|
||||
{
|
||||
name: "Record found in authoritative zone returns normally",
|
||||
zones: []nbdns.CustomZone{{
|
||||
Domain: "custom.zone.",
|
||||
Records: []nbdns.SimpleRecord{record},
|
||||
}},
|
||||
queryName: "existing.custom.zone.",
|
||||
expectFallthrough: false,
|
||||
expectRecord: true,
|
||||
},
|
||||
}
|
||||
|
||||
for _, tc := range testCases {
|
||||
t.Run(tc.name, func(t *testing.T) {
|
||||
resolver.Update(tc.zones)
|
||||
|
||||
var responseMSG *dns.Msg
|
||||
responseWriter := &test.MockResponseWriter{
|
||||
WriteMsgFunc: func(m *dns.Msg) error {
|
||||
responseMSG = m
|
||||
return nil
|
||||
},
|
||||
}
|
||||
|
||||
msg := new(dns.Msg).SetQuestion(tc.queryName, dns.TypeA)
|
||||
resolver.ServeDNS(responseWriter, msg)
|
||||
|
||||
require.NotNil(t, responseMSG, "Should have received a response")
|
||||
|
||||
if tc.expectFallthrough {
|
||||
assert.True(t, responseMSG.MsgHdr.Zero, "Zero bit should be set for fallthrough")
|
||||
assert.Equal(t, dns.RcodeNameError, responseMSG.Rcode, "Should return NXDOMAIN")
|
||||
} else {
|
||||
assert.False(t, responseMSG.MsgHdr.Zero, "Zero bit should not be set")
|
||||
}
|
||||
|
||||
if tc.expectRecord {
|
||||
assert.Greater(t, len(responseMSG.Answer), 0, "Should have answer records")
|
||||
assert.Equal(t, dns.RcodeSuccess, responseMSG.Rcode)
|
||||
}
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
// TestLocalResolver_AuthoritativeFlag tests the AA flag behavior
|
||||
func TestLocalResolver_AuthoritativeFlag(t *testing.T) {
|
||||
t.Run("direct record lookup is authoritative", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
resolver.Update([]nbdns.CustomZone{{
|
||||
Domain: "example.com.",
|
||||
Records: []nbdns.SimpleRecord{
|
||||
{Name: "host.example.com.", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"},
|
||||
},
|
||||
}})
|
||||
|
||||
msg := new(dns.Msg).SetQuestion("host.example.com.", dns.TypeA)
|
||||
var resp *dns.Msg
|
||||
resolver.ServeDNS(&test.MockResponseWriter{WriteMsgFunc: func(m *dns.Msg) error { resp = m; return nil }}, msg)
|
||||
|
||||
require.NotNil(t, resp)
|
||||
assert.True(t, resp.Authoritative)
|
||||
})
|
||||
|
||||
t.Run("external resolution is not authoritative", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
resolver.resolver = &mockResolver{
|
||||
lookupFunc: func(_ context.Context, network, host string) ([]netip.Addr, error) {
|
||||
if host == "external.example.com." && network == "ip4" {
|
||||
return []netip.Addr{netip.MustParseAddr("93.184.216.34")}, nil
|
||||
}
|
||||
return nil, nil
|
||||
},
|
||||
}
|
||||
|
||||
resolver.Update([]nbdns.CustomZone{{
|
||||
Domain: "test.",
|
||||
Records: []nbdns.SimpleRecord{
|
||||
{Name: "alias.test.", Type: int(dns.TypeCNAME), Class: nbdns.DefaultClass, TTL: 300, RData: "external.example.com."},
|
||||
},
|
||||
}})
|
||||
|
||||
msg := new(dns.Msg).SetQuestion("alias.test.", dns.TypeA)
|
||||
var resp *dns.Msg
|
||||
resolver.ServeDNS(&test.MockResponseWriter{WriteMsgFunc: func(m *dns.Msg) error { resp = m; return nil }}, msg)
|
||||
|
||||
require.NotNil(t, resp)
|
||||
require.Len(t, resp.Answer, 2)
|
||||
assert.False(t, resp.Authoritative)
|
||||
})
|
||||
}
|
||||
|
||||
// TestLocalResolver_Stop tests cleanup on Stop
|
||||
func TestLocalResolver_Stop(t *testing.T) {
|
||||
t.Run("Stop clears all state", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
resolver.Update([]nbdns.CustomZone{{
|
||||
Domain: "example.com.",
|
||||
Records: []nbdns.SimpleRecord{
|
||||
{Name: "host.example.com.", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"},
|
||||
},
|
||||
}})
|
||||
|
||||
resolver.Stop()
|
||||
|
||||
msg := new(dns.Msg).SetQuestion("host.example.com.", dns.TypeA)
|
||||
var resp *dns.Msg
|
||||
resolver.ServeDNS(&test.MockResponseWriter{WriteMsgFunc: func(m *dns.Msg) error { resp = m; return nil }}, msg)
|
||||
|
||||
require.NotNil(t, resp)
|
||||
assert.Len(t, resp.Answer, 0)
|
||||
assert.False(t, resolver.isInManagedZone("host.example.com."))
|
||||
})
|
||||
|
||||
t.Run("Stop is safe to call multiple times", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
resolver.Update([]nbdns.CustomZone{{
|
||||
Domain: "example.com.",
|
||||
Records: []nbdns.SimpleRecord{
|
||||
{Name: "host.example.com.", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"},
|
||||
},
|
||||
}})
|
||||
|
||||
resolver.Stop()
|
||||
resolver.Stop()
|
||||
resolver.Stop()
|
||||
})
|
||||
|
||||
t.Run("Stop cancels in-flight external resolution", func(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
|
||||
lookupStarted := make(chan struct{})
|
||||
lookupCtxCanceled := make(chan struct{})
|
||||
|
||||
resolver.resolver = &mockResolver{
|
||||
lookupFunc: func(ctx context.Context, network, host string) ([]netip.Addr, error) {
|
||||
close(lookupStarted)
|
||||
<-ctx.Done()
|
||||
close(lookupCtxCanceled)
|
||||
return nil, ctx.Err()
|
||||
},
|
||||
}
|
||||
|
||||
resolver.Update([]nbdns.CustomZone{{
|
||||
Domain: "test.",
|
||||
Records: []nbdns.SimpleRecord{
|
||||
{Name: "alias.test.", Type: int(dns.TypeCNAME), Class: nbdns.DefaultClass, TTL: 300, RData: "external.example.com."},
|
||||
},
|
||||
}})
|
||||
|
||||
done := make(chan struct{})
|
||||
go func() {
|
||||
msg := new(dns.Msg).SetQuestion("alias.test.", dns.TypeA)
|
||||
resolver.ServeDNS(&test.MockResponseWriter{WriteMsgFunc: func(m *dns.Msg) error { return nil }}, msg)
|
||||
close(done)
|
||||
}()
|
||||
|
||||
<-lookupStarted
|
||||
resolver.Stop()
|
||||
|
||||
select {
|
||||
case <-lookupCtxCanceled:
|
||||
case <-time.After(time.Second):
|
||||
t.Fatal("external lookup context was not canceled")
|
||||
}
|
||||
|
||||
select {
|
||||
case <-done:
|
||||
case <-time.After(time.Second):
|
||||
t.Fatal("ServeDNS did not return after Stop")
|
||||
}
|
||||
})
|
||||
}
|
||||
|
||||
// TestLocalResolver_FallthroughCaseInsensitive verifies case-insensitive domain matching for fallthrough
|
||||
func TestLocalResolver_FallthroughCaseInsensitive(t *testing.T) {
|
||||
resolver := NewResolver()
|
||||
|
||||
resolver.Update([]nbdns.CustomZone{{
|
||||
Domain: "EXAMPLE.COM.",
|
||||
Records: []nbdns.SimpleRecord{{Name: "host.example.com.", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "1.2.3.4"}},
|
||||
NonAuthoritative: true,
|
||||
}})
|
||||
|
||||
var responseMSG *dns.Msg
|
||||
responseWriter := &test.MockResponseWriter{
|
||||
WriteMsgFunc: func(m *dns.Msg) error {
|
||||
responseMSG = m
|
||||
return nil
|
||||
},
|
||||
}
|
||||
|
||||
msg := new(dns.Msg).SetQuestion("nonexistent.example.com.", dns.TypeA)
|
||||
resolver.ServeDNS(responseWriter, msg)
|
||||
|
||||
require.NotNil(t, responseMSG)
|
||||
assert.True(t, responseMSG.MsgHdr.Zero, "Should fallthrough for non-authoritative zone with case-insensitive match")
|
||||
}
|
||||
|
||||
// BenchmarkFindZone_BestCase benchmarks zone lookup with immediate match (first label)
|
||||
func BenchmarkFindZone_BestCase(b *testing.B) {
|
||||
resolver := NewResolver()
|
||||
|
||||
// Single zone that matches immediately
|
||||
resolver.Update([]nbdns.CustomZone{{
|
||||
Domain: "example.com.",
|
||||
NonAuthoritative: true,
|
||||
}})
|
||||
|
||||
b.ResetTimer()
|
||||
for i := 0; i < b.N; i++ {
|
||||
resolver.shouldFallthrough("example.com.")
|
||||
}
|
||||
}
|
||||
|
||||
// BenchmarkFindZone_WorstCase benchmarks zone lookup with many zones, no match, many labels
|
||||
func BenchmarkFindZone_WorstCase(b *testing.B) {
|
||||
resolver := NewResolver()
|
||||
|
||||
// 100 zones that won't match
|
||||
var zones []nbdns.CustomZone
|
||||
for i := 0; i < 100; i++ {
|
||||
zones = append(zones, nbdns.CustomZone{
|
||||
Domain: fmt.Sprintf("zone%d.internal.", i),
|
||||
NonAuthoritative: true,
|
||||
})
|
||||
}
|
||||
resolver.Update(zones)
|
||||
|
||||
// Query with many labels that won't match any zone
|
||||
qname := "a.b.c.d.e.f.g.h.external.com."
|
||||
|
||||
b.ResetTimer()
|
||||
for i := 0; i < b.N; i++ {
|
||||
resolver.shouldFallthrough(qname)
|
||||
}
|
||||
}
|
||||
|
||||
// BenchmarkFindZone_TypicalCase benchmarks typical usage: few zones, subdomain match
|
||||
func BenchmarkFindZone_TypicalCase(b *testing.B) {
|
||||
resolver := NewResolver()
|
||||
|
||||
// Typical setup: peer zone (authoritative) + one user zone (non-authoritative)
|
||||
resolver.Update([]nbdns.CustomZone{
|
||||
{Domain: "netbird.cloud.", NonAuthoritative: false},
|
||||
{Domain: "custom.local.", NonAuthoritative: true},
|
||||
})
|
||||
|
||||
// Query for subdomain of user zone
|
||||
qname := "myhost.custom.local."
|
||||
|
||||
b.ResetTimer()
|
||||
for i := 0; i < b.N; i++ {
|
||||
resolver.shouldFallthrough(qname)
|
||||
}
|
||||
}
|
||||
|
||||
// BenchmarkIsInManagedZone_ManyZones benchmarks isInManagedZone with 100 zones
|
||||
func BenchmarkIsInManagedZone_ManyZones(b *testing.B) {
|
||||
resolver := NewResolver()
|
||||
|
||||
var zones []nbdns.CustomZone
|
||||
for i := 0; i < 100; i++ {
|
||||
zones = append(zones, nbdns.CustomZone{
|
||||
Domain: fmt.Sprintf("zone%d.internal.", i),
|
||||
})
|
||||
}
|
||||
resolver.Update(zones)
|
||||
|
||||
// Query that matches zone50
|
||||
qname := "host.zone50.internal."
|
||||
|
||||
b.ResetTimer()
|
||||
for i := 0; i < b.N; i++ {
|
||||
resolver.isInManagedZone(qname)
|
||||
}
|
||||
}
|
||||
|
||||
@@ -4,7 +4,6 @@ import (
|
||||
"context"
|
||||
"fmt"
|
||||
"net"
|
||||
"net/netip"
|
||||
"net/url"
|
||||
"strings"
|
||||
"sync"
|
||||
@@ -27,11 +26,6 @@ type Resolver struct {
|
||||
mutex sync.RWMutex
|
||||
}
|
||||
|
||||
type ipsResponse struct {
|
||||
ips []netip.Addr
|
||||
err error
|
||||
}
|
||||
|
||||
// NewResolver creates a new management domains cache resolver.
|
||||
func NewResolver() *Resolver {
|
||||
return &Resolver{
|
||||
@@ -105,9 +99,9 @@ func (m *Resolver) AddDomain(ctx context.Context, d domain.Domain) error {
|
||||
ctx, cancel := context.WithTimeout(ctx, dnsTimeout)
|
||||
defer cancel()
|
||||
|
||||
ips, err := lookupIPWithExtraTimeout(ctx, d)
|
||||
ips, err := net.DefaultResolver.LookupNetIP(ctx, "ip", d.PunycodeString())
|
||||
if err != nil {
|
||||
return err
|
||||
return fmt.Errorf("resolve domain %s: %w", d.SafeString(), err)
|
||||
}
|
||||
|
||||
var aRecords, aaaaRecords []dns.RR
|
||||
@@ -165,36 +159,6 @@ func (m *Resolver) AddDomain(ctx context.Context, d domain.Domain) error {
|
||||
return nil
|
||||
}
|
||||
|
||||
func lookupIPWithExtraTimeout(ctx context.Context, d domain.Domain) ([]netip.Addr, error) {
|
||||
log.Infof("looking up IP for mgmt domain=%s", d.SafeString())
|
||||
defer log.Infof("done looking up IP for mgmt domain=%s", d.SafeString())
|
||||
resultChan := make(chan *ipsResponse, 1)
|
||||
|
||||
go func() {
|
||||
ips, err := net.DefaultResolver.LookupNetIP(ctx, "ip", d.PunycodeString())
|
||||
resultChan <- &ipsResponse{
|
||||
err: err,
|
||||
ips: ips,
|
||||
}
|
||||
}()
|
||||
|
||||
var resp *ipsResponse
|
||||
|
||||
select {
|
||||
case <-time.After(dnsTimeout + time.Millisecond*500):
|
||||
log.Warnf("timed out waiting for IP for mgmt domain=%s", d.SafeString())
|
||||
return nil, fmt.Errorf("timed out waiting for ips to be available for domain %s", d.SafeString())
|
||||
case <-ctx.Done():
|
||||
return nil, ctx.Err()
|
||||
case resp = <-resultChan:
|
||||
}
|
||||
|
||||
if resp.err != nil {
|
||||
return nil, fmt.Errorf("resolve domain %s: %w", d.SafeString(), resp.err)
|
||||
}
|
||||
return resp.ips, nil
|
||||
}
|
||||
|
||||
// PopulateFromConfig extracts and caches domains from the client configuration.
|
||||
func (m *Resolver) PopulateFromConfig(ctx context.Context, mgmtURL *url.URL) error {
|
||||
if mgmtURL == nil {
|
||||
|
||||
@@ -1,197 +0,0 @@
|
||||
// Package resutil provides shared DNS resolution utilities
|
||||
package resutil
|
||||
|
||||
import (
|
||||
"context"
|
||||
"crypto/rand"
|
||||
"encoding/hex"
|
||||
"errors"
|
||||
"net"
|
||||
"net/netip"
|
||||
"strings"
|
||||
|
||||
"github.com/miekg/dns"
|
||||
log "github.com/sirupsen/logrus"
|
||||
)
|
||||
|
||||
// GenerateRequestID creates a random 8-character hex string for request tracing.
|
||||
func GenerateRequestID() string {
|
||||
bytes := make([]byte, 4)
|
||||
if _, err := rand.Read(bytes); err != nil {
|
||||
log.Errorf("generate request ID: %v", err)
|
||||
return ""
|
||||
}
|
||||
return hex.EncodeToString(bytes)
|
||||
}
|
||||
|
||||
// IPsToRRs converts a slice of IP addresses to DNS resource records.
|
||||
// IPv4 addresses become A records, IPv6 addresses become AAAA records.
|
||||
func IPsToRRs(name string, ips []netip.Addr, ttl uint32) []dns.RR {
|
||||
var result []dns.RR
|
||||
|
||||
for _, ip := range ips {
|
||||
if ip.Is6() {
|
||||
result = append(result, &dns.AAAA{
|
||||
Hdr: dns.RR_Header{
|
||||
Name: name,
|
||||
Rrtype: dns.TypeAAAA,
|
||||
Class: dns.ClassINET,
|
||||
Ttl: ttl,
|
||||
},
|
||||
AAAA: ip.AsSlice(),
|
||||
})
|
||||
} else {
|
||||
result = append(result, &dns.A{
|
||||
Hdr: dns.RR_Header{
|
||||
Name: name,
|
||||
Rrtype: dns.TypeA,
|
||||
Class: dns.ClassINET,
|
||||
Ttl: ttl,
|
||||
},
|
||||
A: ip.AsSlice(),
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
return result
|
||||
}
|
||||
|
||||
// NetworkForQtype returns the network string ("ip4" or "ip6") for a DNS query type.
|
||||
// Returns empty string for unsupported types.
|
||||
func NetworkForQtype(qtype uint16) string {
|
||||
switch qtype {
|
||||
case dns.TypeA:
|
||||
return "ip4"
|
||||
case dns.TypeAAAA:
|
||||
return "ip6"
|
||||
default:
|
||||
return ""
|
||||
}
|
||||
}
|
||||
|
||||
type resolver interface {
|
||||
LookupNetIP(ctx context.Context, network, host string) ([]netip.Addr, error)
|
||||
}
|
||||
|
||||
// chainedWriter is implemented by ResponseWriters that carry request metadata
|
||||
type chainedWriter interface {
|
||||
RequestID() string
|
||||
SetMeta(key, value string)
|
||||
}
|
||||
|
||||
// GetRequestID extracts a request ID from the ResponseWriter if available,
|
||||
// otherwise generates a new one.
|
||||
func GetRequestID(w dns.ResponseWriter) string {
|
||||
if cw, ok := w.(chainedWriter); ok {
|
||||
if id := cw.RequestID(); id != "" {
|
||||
return id
|
||||
}
|
||||
}
|
||||
return GenerateRequestID()
|
||||
}
|
||||
|
||||
// SetMeta sets metadata on the ResponseWriter if it supports it.
|
||||
func SetMeta(w dns.ResponseWriter, key, value string) {
|
||||
if cw, ok := w.(chainedWriter); ok {
|
||||
cw.SetMeta(key, value)
|
||||
}
|
||||
}
|
||||
|
||||
// LookupResult contains the result of an external DNS lookup
|
||||
type LookupResult struct {
|
||||
IPs []netip.Addr
|
||||
Rcode int
|
||||
Err error // Original error for caller's logging needs
|
||||
}
|
||||
|
||||
// LookupIP performs a DNS lookup and determines the appropriate rcode.
|
||||
func LookupIP(ctx context.Context, r resolver, network, host string, qtype uint16) LookupResult {
|
||||
ips, err := r.LookupNetIP(ctx, network, host)
|
||||
if err != nil {
|
||||
return LookupResult{
|
||||
Rcode: getRcodeForError(ctx, r, host, qtype, err),
|
||||
Err: err,
|
||||
}
|
||||
}
|
||||
|
||||
// Unmap IPv4-mapped IPv6 addresses that some resolvers may return
|
||||
for i, ip := range ips {
|
||||
ips[i] = ip.Unmap()
|
||||
}
|
||||
|
||||
return LookupResult{
|
||||
IPs: ips,
|
||||
Rcode: dns.RcodeSuccess,
|
||||
}
|
||||
}
|
||||
|
||||
func getRcodeForError(ctx context.Context, r resolver, host string, qtype uint16, err error) int {
|
||||
var dnsErr *net.DNSError
|
||||
if !errors.As(err, &dnsErr) {
|
||||
return dns.RcodeServerFailure
|
||||
}
|
||||
|
||||
if dnsErr.IsNotFound {
|
||||
return getRcodeForNotFound(ctx, r, host, qtype)
|
||||
}
|
||||
|
||||
return dns.RcodeServerFailure
|
||||
}
|
||||
|
||||
// getRcodeForNotFound distinguishes between NXDOMAIN (domain doesn't exist) and NODATA
|
||||
// (domain exists but no records of requested type) by checking the opposite record type.
|
||||
//
|
||||
// musl libc (the reason we need this distinction) only queries A/AAAA pairs in getaddrinfo,
|
||||
// so checking the opposite A/AAAA type is sufficient. Other record types (MX, TXT, etc.)
|
||||
// are not queried by musl and don't need this handling.
|
||||
func getRcodeForNotFound(ctx context.Context, r resolver, domain string, originalQtype uint16) int {
|
||||
// Try querying for a different record type to see if the domain exists
|
||||
// If the original query was for AAAA, try A. If it was for A, try AAAA.
|
||||
// This helps distinguish between NXDOMAIN and NODATA.
|
||||
var alternativeNetwork string
|
||||
switch originalQtype {
|
||||
case dns.TypeAAAA:
|
||||
alternativeNetwork = "ip4"
|
||||
case dns.TypeA:
|
||||
alternativeNetwork = "ip6"
|
||||
default:
|
||||
return dns.RcodeNameError
|
||||
}
|
||||
|
||||
if _, err := r.LookupNetIP(ctx, alternativeNetwork, domain); err != nil {
|
||||
var dnsErr *net.DNSError
|
||||
if errors.As(err, &dnsErr) && dnsErr.IsNotFound {
|
||||
// Alternative query also returned not found - domain truly doesn't exist
|
||||
return dns.RcodeNameError
|
||||
}
|
||||
// Some other error (timeout, server failure, etc.) - can't determine, assume domain exists
|
||||
return dns.RcodeSuccess
|
||||
}
|
||||
|
||||
// Alternative query succeeded - domain exists but has no records of this type
|
||||
return dns.RcodeSuccess
|
||||
}
|
||||
|
||||
// FormatAnswers formats DNS resource records for logging.
|
||||
func FormatAnswers(answers []dns.RR) string {
|
||||
if len(answers) == 0 {
|
||||
return "[]"
|
||||
}
|
||||
|
||||
parts := make([]string, 0, len(answers))
|
||||
for _, rr := range answers {
|
||||
switch r := rr.(type) {
|
||||
case *dns.A:
|
||||
parts = append(parts, r.A.String())
|
||||
case *dns.AAAA:
|
||||
parts = append(parts, r.AAAA.String())
|
||||
case *dns.CNAME:
|
||||
parts = append(parts, "CNAME:"+r.Target)
|
||||
case *dns.PTR:
|
||||
parts = append(parts, "PTR:"+r.Ptr)
|
||||
default:
|
||||
parts = append(parts, dns.TypeToString[rr.Header().Rrtype])
|
||||
}
|
||||
}
|
||||
return "[" + strings.Join(parts, ", ") + "]"
|
||||
}
|
||||
@@ -80,7 +80,6 @@ type DefaultServer struct {
|
||||
updateSerial uint64
|
||||
previousConfigHash uint64
|
||||
currentConfig HostDNSConfig
|
||||
currentConfigHash uint64
|
||||
handlerChain *HandlerChain
|
||||
extraDomains map[domain.Domain]int
|
||||
|
||||
@@ -208,7 +207,6 @@ func newDefaultServer(
|
||||
hostsDNSHolder: newHostsDNSHolder(),
|
||||
hostManager: &noopHostConfigurator{},
|
||||
mgmtCacheResolver: mgmtCacheResolver,
|
||||
currentConfigHash: ^uint64(0), // Initialize to max uint64 to ensure first config is always applied
|
||||
}
|
||||
|
||||
// register with root zone, handler chain takes care of the routing
|
||||
@@ -485,7 +483,7 @@ func (s *DefaultServer) applyConfiguration(update nbdns.Config) error {
|
||||
}
|
||||
}
|
||||
|
||||
localMuxUpdates, localZones, err := s.buildLocalHandlerUpdate(update.CustomZones)
|
||||
localMuxUpdates, localRecords, err := s.buildLocalHandlerUpdate(update.CustomZones)
|
||||
if err != nil {
|
||||
return fmt.Errorf("local handler updater: %w", err)
|
||||
}
|
||||
@@ -498,7 +496,8 @@ func (s *DefaultServer) applyConfiguration(update nbdns.Config) error {
|
||||
|
||||
s.updateMux(muxUpdates)
|
||||
|
||||
s.localResolver.Update(localZones)
|
||||
// register local records
|
||||
s.localResolver.Update(localRecords)
|
||||
|
||||
s.currentConfig = dnsConfigToHostDNSConfig(update, s.service.RuntimeIP(), s.service.RuntimePort())
|
||||
|
||||
@@ -587,29 +586,8 @@ func (s *DefaultServer) applyHostConfig() {
|
||||
|
||||
log.Debugf("extra match domains: %v", maps.Keys(s.extraDomains))
|
||||
|
||||
hash, err := hashstructure.Hash(config, hashstructure.FormatV2, &hashstructure.HashOptions{
|
||||
ZeroNil: true,
|
||||
IgnoreZeroValue: true,
|
||||
SlicesAsSets: true,
|
||||
UseStringer: true,
|
||||
})
|
||||
if err != nil {
|
||||
log.Warnf("unable to hash the host dns configuration, will apply config anyway: %s", err)
|
||||
// Fall through to apply config anyway (fail-safe approach)
|
||||
} else if s.currentConfigHash == hash {
|
||||
log.Debugf("not applying host config as there are no changes")
|
||||
return
|
||||
}
|
||||
|
||||
log.Debugf("applying host config as there are changes")
|
||||
if err := s.hostManager.applyDNSConfig(config, s.stateManager); err != nil {
|
||||
log.Errorf("failed to apply DNS host manager update: %v", err)
|
||||
return
|
||||
}
|
||||
|
||||
// Only update hash if it was computed successfully and config was applied
|
||||
if err == nil {
|
||||
s.currentConfigHash = hash
|
||||
}
|
||||
|
||||
s.registerFallback(config)
|
||||
@@ -631,7 +609,9 @@ func (s *DefaultServer) registerFallback(config HostDNSConfig) {
|
||||
|
||||
handler, err := newUpstreamResolver(
|
||||
s.ctx,
|
||||
s.wgInterface,
|
||||
s.wgInterface.Name(),
|
||||
s.wgInterface.Address().IP,
|
||||
s.wgInterface.Address().Network,
|
||||
s.statusRecorder,
|
||||
s.hostsDNSHolder,
|
||||
nbdns.RootZone,
|
||||
@@ -656,9 +636,9 @@ func (s *DefaultServer) registerFallback(config HostDNSConfig) {
|
||||
s.registerHandler([]string{nbdns.RootZone}, handler, PriorityFallback)
|
||||
}
|
||||
|
||||
func (s *DefaultServer) buildLocalHandlerUpdate(customZones []nbdns.CustomZone) ([]handlerWrapper, []nbdns.CustomZone, error) {
|
||||
func (s *DefaultServer) buildLocalHandlerUpdate(customZones []nbdns.CustomZone) ([]handlerWrapper, []nbdns.SimpleRecord, error) {
|
||||
var muxUpdates []handlerWrapper
|
||||
var zones []nbdns.CustomZone
|
||||
var localRecords []nbdns.SimpleRecord
|
||||
|
||||
for _, customZone := range customZones {
|
||||
if len(customZone.Records) == 0 {
|
||||
@@ -672,20 +652,17 @@ func (s *DefaultServer) buildLocalHandlerUpdate(customZones []nbdns.CustomZone)
|
||||
priority: PriorityLocal,
|
||||
})
|
||||
|
||||
// zone records contain the fqdn, so we can just flatten them
|
||||
var localRecords []nbdns.SimpleRecord
|
||||
for _, record := range customZone.Records {
|
||||
if record.Class != nbdns.DefaultClass {
|
||||
log.Warnf("received an invalid class type: %s", record.Class)
|
||||
continue
|
||||
}
|
||||
// zone records contain the fqdn, so we can just flatten them
|
||||
localRecords = append(localRecords, record)
|
||||
}
|
||||
customZone.Records = localRecords
|
||||
zones = append(zones, customZone)
|
||||
}
|
||||
|
||||
return muxUpdates, zones, nil
|
||||
return muxUpdates, localRecords, nil
|
||||
}
|
||||
|
||||
func (s *DefaultServer) buildUpstreamHandlerUpdate(nameServerGroups []*nbdns.NameServerGroup) ([]handlerWrapper, error) {
|
||||
@@ -741,7 +718,9 @@ func (s *DefaultServer) createHandlersForDomainGroup(domainGroup nsGroupsByDomai
|
||||
log.Debugf("creating handler for domain=%s with priority=%d", domainGroup.domain, priority)
|
||||
handler, err := newUpstreamResolver(
|
||||
s.ctx,
|
||||
s.wgInterface,
|
||||
s.wgInterface.Name(),
|
||||
s.wgInterface.Address().IP,
|
||||
s.wgInterface.Address().Network,
|
||||
s.statusRecorder,
|
||||
s.hostsDNSHolder,
|
||||
domainGroup.domain,
|
||||
@@ -922,7 +901,9 @@ func (s *DefaultServer) addHostRootZone() {
|
||||
|
||||
handler, err := newUpstreamResolver(
|
||||
s.ctx,
|
||||
s.wgInterface,
|
||||
s.wgInterface.Name(),
|
||||
s.wgInterface.Address().IP,
|
||||
s.wgInterface.Address().Network,
|
||||
s.statusRecorder,
|
||||
s.hostsDNSHolder,
|
||||
nbdns.RootZone,
|
||||
|
||||
@@ -15,7 +15,6 @@ import (
|
||||
log "github.com/sirupsen/logrus"
|
||||
"github.com/stretchr/testify/assert"
|
||||
"github.com/stretchr/testify/mock"
|
||||
"golang.zx2c4.com/wireguard/tun/netstack"
|
||||
"golang.zx2c4.com/wireguard/wgctrl/wgtypes"
|
||||
|
||||
"github.com/netbirdio/netbird/client/firewall/uspfilter"
|
||||
@@ -82,10 +81,6 @@ func (w *mocWGIface) GetStats(_ string) (configurer.WGStats, error) {
|
||||
return configurer.WGStats{}, nil
|
||||
}
|
||||
|
||||
func (w *mocWGIface) GetNet() *netstack.Net {
|
||||
return nil
|
||||
}
|
||||
|
||||
var zoneRecords = []nbdns.SimpleRecord{
|
||||
{
|
||||
Name: "peera.netbird.cloud",
|
||||
@@ -133,7 +128,7 @@ func TestUpdateDNSServer(t *testing.T) {
|
||||
testCases := []struct {
|
||||
name string
|
||||
initUpstreamMap registeredHandlerMap
|
||||
initLocalZones []nbdns.CustomZone
|
||||
initLocalRecords []nbdns.SimpleRecord
|
||||
initSerial uint64
|
||||
inputSerial uint64
|
||||
inputUpdate nbdns.Config
|
||||
@@ -185,8 +180,8 @@ func TestUpdateDNSServer(t *testing.T) {
|
||||
expectedLocalQs: []dns.Question{{Name: "peera.netbird.cloud.", Qtype: dns.TypeA, Qclass: dns.ClassINET}},
|
||||
},
|
||||
{
|
||||
name: "New Config Should Succeed",
|
||||
initLocalZones: []nbdns.CustomZone{{Domain: "netbird.cloud", Records: []nbdns.SimpleRecord{{Name: "netbird.cloud", Type: 1, Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"}}}},
|
||||
name: "New Config Should Succeed",
|
||||
initLocalRecords: []nbdns.SimpleRecord{{Name: "netbird.cloud", Type: 1, Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"}},
|
||||
initUpstreamMap: registeredHandlerMap{
|
||||
generateDummyHandler(zoneRecords[0].Name, nameServers).ID(): handlerWrapper{
|
||||
domain: "netbird.cloud",
|
||||
@@ -226,19 +221,19 @@ func TestUpdateDNSServer(t *testing.T) {
|
||||
expectedLocalQs: []dns.Question{{Name: zoneRecords[0].Name, Qtype: 1, Qclass: 1}},
|
||||
},
|
||||
{
|
||||
name: "Smaller Config Serial Should Be Skipped",
|
||||
initLocalZones: []nbdns.CustomZone{},
|
||||
initUpstreamMap: make(registeredHandlerMap),
|
||||
initSerial: 2,
|
||||
inputSerial: 1,
|
||||
shouldFail: true,
|
||||
name: "Smaller Config Serial Should Be Skipped",
|
||||
initLocalRecords: []nbdns.SimpleRecord{},
|
||||
initUpstreamMap: make(registeredHandlerMap),
|
||||
initSerial: 2,
|
||||
inputSerial: 1,
|
||||
shouldFail: true,
|
||||
},
|
||||
{
|
||||
name: "Empty NS Group Domain Or Not Primary Element Should Fail",
|
||||
initLocalZones: []nbdns.CustomZone{},
|
||||
initUpstreamMap: make(registeredHandlerMap),
|
||||
initSerial: 0,
|
||||
inputSerial: 1,
|
||||
name: "Empty NS Group Domain Or Not Primary Element Should Fail",
|
||||
initLocalRecords: []nbdns.SimpleRecord{},
|
||||
initUpstreamMap: make(registeredHandlerMap),
|
||||
initSerial: 0,
|
||||
inputSerial: 1,
|
||||
inputUpdate: nbdns.Config{
|
||||
ServiceEnable: true,
|
||||
CustomZones: []nbdns.CustomZone{
|
||||
@@ -256,11 +251,11 @@ func TestUpdateDNSServer(t *testing.T) {
|
||||
shouldFail: true,
|
||||
},
|
||||
{
|
||||
name: "Invalid NS Group Nameservers list Should Fail",
|
||||
initLocalZones: []nbdns.CustomZone{},
|
||||
initUpstreamMap: make(registeredHandlerMap),
|
||||
initSerial: 0,
|
||||
inputSerial: 1,
|
||||
name: "Invalid NS Group Nameservers list Should Fail",
|
||||
initLocalRecords: []nbdns.SimpleRecord{},
|
||||
initUpstreamMap: make(registeredHandlerMap),
|
||||
initSerial: 0,
|
||||
inputSerial: 1,
|
||||
inputUpdate: nbdns.Config{
|
||||
ServiceEnable: true,
|
||||
CustomZones: []nbdns.CustomZone{
|
||||
@@ -278,11 +273,11 @@ func TestUpdateDNSServer(t *testing.T) {
|
||||
shouldFail: true,
|
||||
},
|
||||
{
|
||||
name: "Invalid Custom Zone Records list Should Skip",
|
||||
initLocalZones: []nbdns.CustomZone{},
|
||||
initUpstreamMap: make(registeredHandlerMap),
|
||||
initSerial: 0,
|
||||
inputSerial: 1,
|
||||
name: "Invalid Custom Zone Records list Should Skip",
|
||||
initLocalRecords: []nbdns.SimpleRecord{},
|
||||
initUpstreamMap: make(registeredHandlerMap),
|
||||
initSerial: 0,
|
||||
inputSerial: 1,
|
||||
inputUpdate: nbdns.Config{
|
||||
ServiceEnable: true,
|
||||
CustomZones: []nbdns.CustomZone{
|
||||
@@ -304,8 +299,8 @@ func TestUpdateDNSServer(t *testing.T) {
|
||||
}},
|
||||
},
|
||||
{
|
||||
name: "Empty Config Should Succeed and Clean Maps",
|
||||
initLocalZones: []nbdns.CustomZone{{Domain: "netbird.cloud", Records: []nbdns.SimpleRecord{{Name: "netbird.cloud", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"}}}},
|
||||
name: "Empty Config Should Succeed and Clean Maps",
|
||||
initLocalRecords: []nbdns.SimpleRecord{{Name: "netbird.cloud", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"}},
|
||||
initUpstreamMap: registeredHandlerMap{
|
||||
generateDummyHandler(zoneRecords[0].Name, nameServers).ID(): handlerWrapper{
|
||||
domain: zoneRecords[0].Name,
|
||||
@@ -320,8 +315,8 @@ func TestUpdateDNSServer(t *testing.T) {
|
||||
expectedLocalQs: []dns.Question{},
|
||||
},
|
||||
{
|
||||
name: "Disabled Service Should clean map",
|
||||
initLocalZones: []nbdns.CustomZone{{Domain: "netbird.cloud", Records: []nbdns.SimpleRecord{{Name: "netbird.cloud", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"}}}},
|
||||
name: "Disabled Service Should clean map",
|
||||
initLocalRecords: []nbdns.SimpleRecord{{Name: "netbird.cloud", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"}},
|
||||
initUpstreamMap: registeredHandlerMap{
|
||||
generateDummyHandler(zoneRecords[0].Name, nameServers).ID(): handlerWrapper{
|
||||
domain: zoneRecords[0].Name,
|
||||
@@ -390,7 +385,7 @@ func TestUpdateDNSServer(t *testing.T) {
|
||||
}()
|
||||
|
||||
dnsServer.dnsMuxMap = testCase.initUpstreamMap
|
||||
dnsServer.localResolver.Update(testCase.initLocalZones)
|
||||
dnsServer.localResolver.Update(testCase.initLocalRecords)
|
||||
dnsServer.updateSerial = testCase.initSerial
|
||||
|
||||
err = dnsServer.UpdateDNSServer(testCase.inputSerial, testCase.inputUpdate)
|
||||
@@ -515,7 +510,8 @@ func TestDNSFakeResolverHandleUpdates(t *testing.T) {
|
||||
priority: PriorityUpstream,
|
||||
},
|
||||
}
|
||||
dnsServer.localResolver.Update([]nbdns.CustomZone{{Domain: "netbird.cloud", Records: []nbdns.SimpleRecord{{Name: "netbird.cloud", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"}}}})
|
||||
//dnsServer.localResolver.RegisteredMap = local.RegistrationMap{local.BuildRecordKey("netbird.cloud", dns.ClassINET, dns.TypeA): struct{}{}}
|
||||
dnsServer.localResolver.Update([]nbdns.SimpleRecord{{Name: "netbird.cloud", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"}})
|
||||
dnsServer.updateSerial = 0
|
||||
|
||||
nameServers := []nbdns.NameServer{
|
||||
@@ -1606,10 +1602,7 @@ func TestExtraDomains(t *testing.T) {
|
||||
"other.example.com.",
|
||||
"duplicate.example.com.",
|
||||
},
|
||||
// Expect 3 calls instead of 4 because when deregistering duplicate.example.com,
|
||||
// the domain remains in the config (ref count goes from 2 to 1), so the host
|
||||
// config hash doesn't change and applyDNSConfig is not called.
|
||||
applyHostConfigCall: 3,
|
||||
applyHostConfigCall: 4,
|
||||
},
|
||||
{
|
||||
name: "Config update with new domains after registration",
|
||||
@@ -1664,10 +1657,7 @@ func TestExtraDomains(t *testing.T) {
|
||||
expectedMatchOnly: []string{
|
||||
"extra.example.com.",
|
||||
},
|
||||
// Expect 2 calls instead of 3 because when deregistering protected.example.com,
|
||||
// it's removed from extraDomains but still remains in the config (from customZones),
|
||||
// so the host config hash doesn't change and applyDNSConfig is not called.
|
||||
applyHostConfigCall: 2,
|
||||
applyHostConfigCall: 3,
|
||||
},
|
||||
{
|
||||
name: "Register domain that is part of nameserver group",
|
||||
@@ -2052,7 +2042,7 @@ func TestLocalResolverPriorityInServer(t *testing.T) {
|
||||
|
||||
func TestLocalResolverPriorityConstants(t *testing.T) {
|
||||
// Test that priority constants are ordered correctly
|
||||
assert.Greater(t, PriorityDNSRoute, PriorityLocal, "DNS Route should be higher than Local priority")
|
||||
assert.Greater(t, PriorityLocal, PriorityDNSRoute, "Local priority should be higher than DNS route")
|
||||
assert.Greater(t, PriorityLocal, PriorityUpstream, "Local priority should be higher than upstream")
|
||||
assert.Greater(t, PriorityUpstream, PriorityDefault, "Upstream priority should be higher than default")
|
||||
|
||||
|
||||
@@ -2,6 +2,7 @@ package dns
|
||||
|
||||
import (
|
||||
"context"
|
||||
"crypto/rand"
|
||||
"crypto/sha256"
|
||||
"encoding/hex"
|
||||
"errors"
|
||||
@@ -18,10 +19,8 @@ import (
|
||||
"github.com/hashicorp/go-multierror"
|
||||
"github.com/miekg/dns"
|
||||
log "github.com/sirupsen/logrus"
|
||||
"golang.zx2c4.com/wireguard/tun/netstack"
|
||||
|
||||
"github.com/netbirdio/netbird/client/iface"
|
||||
"github.com/netbirdio/netbird/client/internal/dns/resutil"
|
||||
"github.com/netbirdio/netbird/client/internal/dns/types"
|
||||
"github.com/netbirdio/netbird/client/internal/peer"
|
||||
"github.com/netbirdio/netbird/client/proto"
|
||||
@@ -114,7 +113,10 @@ func (u *upstreamResolverBase) Stop() {
|
||||
|
||||
// ServeDNS handles a DNS request
|
||||
func (u *upstreamResolverBase) ServeDNS(w dns.ResponseWriter, r *dns.Msg) {
|
||||
logger := log.WithField("request_id", resutil.GetRequestID(w))
|
||||
requestID := GenerateRequestID()
|
||||
logger := log.WithField("request_id", requestID)
|
||||
|
||||
logger.Tracef("received upstream question: domain=%s type=%v class=%v", r.Question[0].Name, r.Question[0].Qtype, r.Question[0].Qclass)
|
||||
|
||||
u.prepareRequest(r)
|
||||
|
||||
@@ -200,18 +202,11 @@ func (u *upstreamResolverBase) handleUpstreamError(err error, upstream netip.Add
|
||||
|
||||
func (u *upstreamResolverBase) writeSuccessResponse(w dns.ResponseWriter, rm *dns.Msg, upstream netip.AddrPort, domain string, t time.Duration, logger *log.Entry) bool {
|
||||
u.successCount.Add(1)
|
||||
|
||||
resutil.SetMeta(w, "upstream", upstream.String())
|
||||
|
||||
// Clear Zero bit from external responses to prevent upstream servers from
|
||||
// manipulating our internal fallthrough signaling mechanism
|
||||
rm.MsgHdr.Zero = false
|
||||
logger.Tracef("took %s to query the upstream %s for question domain=%s", t, upstream, domain)
|
||||
|
||||
if err := w.WriteMsg(rm); err != nil {
|
||||
logger.Errorf("failed to write DNS response for question domain=%s: %s", domain, err)
|
||||
return true
|
||||
}
|
||||
|
||||
return true
|
||||
}
|
||||
|
||||
@@ -419,56 +414,16 @@ func ExchangeWithFallback(ctx context.Context, client *dns.Client, r *dns.Msg, u
|
||||
return rm, t, nil
|
||||
}
|
||||
|
||||
// ExchangeWithNetstack performs a DNS exchange using netstack for dialing.
|
||||
// This is needed when netstack is enabled to reach peer IPs through the tunnel.
|
||||
func ExchangeWithNetstack(ctx context.Context, nsNet *netstack.Net, r *dns.Msg, upstream string) (*dns.Msg, error) {
|
||||
reply, err := netstackExchange(ctx, nsNet, r, upstream, "udp")
|
||||
func GenerateRequestID() string {
|
||||
bytes := make([]byte, 4)
|
||||
_, err := rand.Read(bytes)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
log.Errorf("failed to generate request ID: %v", err)
|
||||
return ""
|
||||
}
|
||||
|
||||
// If response is truncated, retry with TCP
|
||||
if reply != nil && reply.MsgHdr.Truncated {
|
||||
log.Tracef("udp response for domain=%s type=%v class=%v is truncated, trying TCP",
|
||||
r.Question[0].Name, r.Question[0].Qtype, r.Question[0].Qclass)
|
||||
return netstackExchange(ctx, nsNet, r, upstream, "tcp")
|
||||
}
|
||||
|
||||
return reply, nil
|
||||
return hex.EncodeToString(bytes)
|
||||
}
|
||||
|
||||
func netstackExchange(ctx context.Context, nsNet *netstack.Net, r *dns.Msg, upstream, network string) (*dns.Msg, error) {
|
||||
conn, err := nsNet.DialContext(ctx, network, upstream)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("with %s: %w", network, err)
|
||||
}
|
||||
defer func() {
|
||||
if err := conn.Close(); err != nil {
|
||||
log.Debugf("failed to close DNS connection: %v", err)
|
||||
}
|
||||
}()
|
||||
|
||||
if deadline, ok := ctx.Deadline(); ok {
|
||||
if err := conn.SetDeadline(deadline); err != nil {
|
||||
return nil, fmt.Errorf("set deadline: %w", err)
|
||||
}
|
||||
}
|
||||
|
||||
dnsConn := &dns.Conn{Conn: conn}
|
||||
|
||||
if err := dnsConn.WriteMsg(r); err != nil {
|
||||
return nil, fmt.Errorf("write %s message: %w", network, err)
|
||||
}
|
||||
|
||||
reply, err := dnsConn.ReadMsg()
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("read %s message: %w", network, err)
|
||||
}
|
||||
|
||||
return reply, nil
|
||||
}
|
||||
|
||||
|
||||
// FormatPeerStatus formats peer connection status information for debugging DNS timeouts
|
||||
func FormatPeerStatus(peerState *peer.State) string {
|
||||
isConnected := peerState.ConnStatus == peer.StatusConnected
|
||||
|
||||
@@ -23,7 +23,9 @@ type upstreamResolver struct {
|
||||
// first time, and we need to wait for a while to start to use again the proper DNS resolver.
|
||||
func newUpstreamResolver(
|
||||
ctx context.Context,
|
||||
_ WGIface,
|
||||
_ string,
|
||||
_ netip.Addr,
|
||||
_ netip.Prefix,
|
||||
statusRecorder *peer.Status,
|
||||
hostsDNSHolder *hostsDNSHolder,
|
||||
domain string,
|
||||
|
||||
@@ -5,23 +5,22 @@ package dns
|
||||
import (
|
||||
"context"
|
||||
"net/netip"
|
||||
"runtime"
|
||||
"time"
|
||||
|
||||
"github.com/miekg/dns"
|
||||
"golang.zx2c4.com/wireguard/tun/netstack"
|
||||
|
||||
"github.com/netbirdio/netbird/client/internal/peer"
|
||||
)
|
||||
|
||||
type upstreamResolver struct {
|
||||
*upstreamResolverBase
|
||||
nsNet *netstack.Net
|
||||
}
|
||||
|
||||
func newUpstreamResolver(
|
||||
ctx context.Context,
|
||||
wgIface WGIface,
|
||||
_ string,
|
||||
_ netip.Addr,
|
||||
_ netip.Prefix,
|
||||
statusRecorder *peer.Status,
|
||||
_ *hostsDNSHolder,
|
||||
domain string,
|
||||
@@ -29,23 +28,12 @@ func newUpstreamResolver(
|
||||
upstreamResolverBase := newUpstreamResolverBase(ctx, statusRecorder, domain)
|
||||
nonIOS := &upstreamResolver{
|
||||
upstreamResolverBase: upstreamResolverBase,
|
||||
nsNet: wgIface.GetNet(),
|
||||
}
|
||||
upstreamResolverBase.upstreamClient = nonIOS
|
||||
return nonIOS, nil
|
||||
}
|
||||
|
||||
func (u *upstreamResolver) exchange(ctx context.Context, upstream string, r *dns.Msg) (rm *dns.Msg, t time.Duration, err error) {
|
||||
// TODO: Check if upstream DNS server is routed through a peer before using netstack.
|
||||
// Similar to iOS logic, we should determine if the DNS server is reachable directly
|
||||
// or needs to go through the tunnel, and only use netstack when necessary.
|
||||
// For now, only use netstack on JS platform where direct access is not possible.
|
||||
if u.nsNet != nil && runtime.GOOS == "js" {
|
||||
start := time.Now()
|
||||
reply, err := ExchangeWithNetstack(ctx, u.nsNet, r, upstream)
|
||||
return reply, time.Since(start), err
|
||||
}
|
||||
|
||||
client := &dns.Client{
|
||||
Timeout: ClientTimeout,
|
||||
}
|
||||
|
||||
@@ -26,7 +26,9 @@ type upstreamResolverIOS struct {
|
||||
|
||||
func newUpstreamResolver(
|
||||
ctx context.Context,
|
||||
wgIface WGIface,
|
||||
interfaceName string,
|
||||
ip netip.Addr,
|
||||
net netip.Prefix,
|
||||
statusRecorder *peer.Status,
|
||||
_ *hostsDNSHolder,
|
||||
domain string,
|
||||
@@ -35,9 +37,9 @@ func newUpstreamResolver(
|
||||
|
||||
ios := &upstreamResolverIOS{
|
||||
upstreamResolverBase: upstreamResolverBase,
|
||||
lIP: wgIface.Address().IP,
|
||||
lNet: wgIface.Address().Network,
|
||||
interfaceName: wgIface.Name(),
|
||||
lIP: ip,
|
||||
lNet: net,
|
||||
interfaceName: interfaceName,
|
||||
}
|
||||
ios.upstreamClient = ios
|
||||
|
||||
|
||||
@@ -2,17 +2,13 @@ package dns
|
||||
|
||||
import (
|
||||
"context"
|
||||
"net"
|
||||
"net/netip"
|
||||
"strings"
|
||||
"testing"
|
||||
"time"
|
||||
|
||||
"github.com/miekg/dns"
|
||||
"golang.zx2c4.com/wireguard/tun/netstack"
|
||||
|
||||
"github.com/netbirdio/netbird/client/iface/device"
|
||||
"github.com/netbirdio/netbird/client/iface/wgaddr"
|
||||
"github.com/netbirdio/netbird/client/internal/dns/test"
|
||||
)
|
||||
|
||||
@@ -62,7 +58,7 @@ func TestUpstreamResolver_ServeDNS(t *testing.T) {
|
||||
for _, testCase := range testCases {
|
||||
t.Run(testCase.name, func(t *testing.T) {
|
||||
ctx, cancel := context.WithCancel(context.TODO())
|
||||
resolver, _ := newUpstreamResolver(ctx, &mockNetstackProvider{}, nil, nil, ".")
|
||||
resolver, _ := newUpstreamResolver(ctx, "", netip.Addr{}, netip.Prefix{}, nil, nil, ".")
|
||||
// Convert test servers to netip.AddrPort
|
||||
var servers []netip.AddrPort
|
||||
for _, server := range testCase.InputServers {
|
||||
@@ -116,19 +112,6 @@ func TestUpstreamResolver_ServeDNS(t *testing.T) {
|
||||
}
|
||||
}
|
||||
|
||||
type mockNetstackProvider struct{}
|
||||
|
||||
func (m *mockNetstackProvider) Name() string { return "mock" }
|
||||
func (m *mockNetstackProvider) Address() wgaddr.Address { return wgaddr.Address{} }
|
||||
func (m *mockNetstackProvider) ToInterface() *net.Interface { return nil }
|
||||
func (m *mockNetstackProvider) IsUserspaceBind() bool { return false }
|
||||
func (m *mockNetstackProvider) GetFilter() device.PacketFilter { return nil }
|
||||
func (m *mockNetstackProvider) GetDevice() *device.FilteredDevice { return nil }
|
||||
func (m *mockNetstackProvider) GetNet() *netstack.Net { return nil }
|
||||
func (m *mockNetstackProvider) GetInterfaceGUIDString() (string, error) {
|
||||
return "", nil
|
||||
}
|
||||
|
||||
type mockUpstreamResolver struct {
|
||||
r *dns.Msg
|
||||
rtt time.Duration
|
||||
|
||||
@@ -5,8 +5,6 @@ package dns
|
||||
import (
|
||||
"net"
|
||||
|
||||
"golang.zx2c4.com/wireguard/tun/netstack"
|
||||
|
||||
"github.com/netbirdio/netbird/client/iface/device"
|
||||
"github.com/netbirdio/netbird/client/iface/wgaddr"
|
||||
)
|
||||
@@ -19,5 +17,4 @@ type WGIface interface {
|
||||
IsUserspaceBind() bool
|
||||
GetFilter() device.PacketFilter
|
||||
GetDevice() *device.FilteredDevice
|
||||
GetNet() *netstack.Net
|
||||
}
|
||||
|
||||
@@ -1,8 +1,6 @@
|
||||
package dns
|
||||
|
||||
import (
|
||||
"golang.zx2c4.com/wireguard/tun/netstack"
|
||||
|
||||
"github.com/netbirdio/netbird/client/iface/device"
|
||||
"github.com/netbirdio/netbird/client/iface/wgaddr"
|
||||
)
|
||||
@@ -14,6 +12,5 @@ type WGIface interface {
|
||||
IsUserspaceBind() bool
|
||||
GetFilter() device.PacketFilter
|
||||
GetDevice() *device.FilteredDevice
|
||||
GetNet() *netstack.Net
|
||||
GetInterfaceGUIDString() (string, error)
|
||||
}
|
||||
|
||||
@@ -18,7 +18,6 @@ import (
|
||||
|
||||
nberrors "github.com/netbirdio/netbird/client/errors"
|
||||
firewall "github.com/netbirdio/netbird/client/firewall/manager"
|
||||
"github.com/netbirdio/netbird/client/internal/dns/resutil"
|
||||
"github.com/netbirdio/netbird/client/internal/peer"
|
||||
"github.com/netbirdio/netbird/route"
|
||||
)
|
||||
@@ -190,22 +189,29 @@ func (f *DNSForwarder) Close(ctx context.Context) error {
|
||||
return nberrors.FormatErrorOrNil(result)
|
||||
}
|
||||
|
||||
func (f *DNSForwarder) handleDNSQuery(logger *log.Entry, w dns.ResponseWriter, query *dns.Msg) *dns.Msg {
|
||||
func (f *DNSForwarder) handleDNSQuery(w dns.ResponseWriter, query *dns.Msg) *dns.Msg {
|
||||
if len(query.Question) == 0 {
|
||||
return nil
|
||||
}
|
||||
question := query.Question[0]
|
||||
logger.Tracef("received DNS request for DNS forwarder: domain=%s type=%s class=%s",
|
||||
question.Name, dns.TypeToString[question.Qtype], dns.ClassToString[question.Qclass])
|
||||
log.Tracef("received DNS request for DNS forwarder: domain=%v type=%v class=%v",
|
||||
question.Name, question.Qtype, question.Qclass)
|
||||
|
||||
domain := strings.ToLower(question.Name)
|
||||
|
||||
resp := query.SetReply(query)
|
||||
network := resutil.NetworkForQtype(question.Qtype)
|
||||
if network == "" {
|
||||
var network string
|
||||
switch question.Qtype {
|
||||
case dns.TypeA:
|
||||
network = "ip4"
|
||||
case dns.TypeAAAA:
|
||||
network = "ip6"
|
||||
default:
|
||||
// TODO: Handle other types
|
||||
|
||||
resp.Rcode = dns.RcodeNotImplemented
|
||||
if err := w.WriteMsg(resp); err != nil {
|
||||
logger.Errorf("failed to write DNS response: %v", err)
|
||||
log.Errorf("failed to write DNS response: %v", err)
|
||||
}
|
||||
return nil
|
||||
}
|
||||
@@ -215,35 +221,33 @@ func (f *DNSForwarder) handleDNSQuery(logger *log.Entry, w dns.ResponseWriter, q
|
||||
if mostSpecificResId == "" {
|
||||
resp.Rcode = dns.RcodeRefused
|
||||
if err := w.WriteMsg(resp); err != nil {
|
||||
logger.Errorf("failed to write DNS response: %v", err)
|
||||
log.Errorf("failed to write DNS response: %v", err)
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
ctx, cancel := context.WithTimeout(context.Background(), upstreamTimeout)
|
||||
defer cancel()
|
||||
|
||||
result := resutil.LookupIP(ctx, f.resolver, network, domain, question.Qtype)
|
||||
if result.Err != nil {
|
||||
f.handleDNSError(ctx, logger, w, question, resp, domain, result)
|
||||
ips, err := f.resolver.LookupNetIP(ctx, network, domain)
|
||||
if err != nil {
|
||||
f.handleDNSError(ctx, w, question, resp, domain, err)
|
||||
return nil
|
||||
}
|
||||
|
||||
f.updateInternalState(result.IPs, mostSpecificResId, matchingEntries)
|
||||
resp.Answer = append(resp.Answer, resutil.IPsToRRs(domain, result.IPs, f.ttl)...)
|
||||
f.cache.set(domain, question.Qtype, result.IPs)
|
||||
// Unmap IPv4-mapped IPv6 addresses that some resolvers may return
|
||||
for i, ip := range ips {
|
||||
ips[i] = ip.Unmap()
|
||||
}
|
||||
|
||||
f.updateInternalState(ips, mostSpecificResId, matchingEntries)
|
||||
f.addIPsToResponse(resp, domain, ips)
|
||||
f.cache.set(domain, question.Qtype, ips)
|
||||
|
||||
return resp
|
||||
}
|
||||
|
||||
func (f *DNSForwarder) handleDNSQueryUDP(w dns.ResponseWriter, query *dns.Msg) {
|
||||
startTime := time.Now()
|
||||
logger := log.WithFields(log.Fields{
|
||||
"request_id": resutil.GenerateRequestID(),
|
||||
"dns_id": fmt.Sprintf("%04x", query.Id),
|
||||
})
|
||||
|
||||
resp := f.handleDNSQuery(logger, w, query)
|
||||
resp := f.handleDNSQuery(w, query)
|
||||
if resp == nil {
|
||||
return
|
||||
}
|
||||
@@ -261,33 +265,19 @@ func (f *DNSForwarder) handleDNSQueryUDP(w dns.ResponseWriter, query *dns.Msg) {
|
||||
}
|
||||
|
||||
if err := w.WriteMsg(resp); err != nil {
|
||||
logger.Errorf("failed to write DNS response: %v", err)
|
||||
return
|
||||
log.Errorf("failed to write DNS response: %v", err)
|
||||
}
|
||||
|
||||
logger.Tracef("response: domain=%s rcode=%s answers=%s took=%s",
|
||||
query.Question[0].Name, dns.RcodeToString[resp.Rcode], resutil.FormatAnswers(resp.Answer), time.Since(startTime))
|
||||
}
|
||||
|
||||
func (f *DNSForwarder) handleDNSQueryTCP(w dns.ResponseWriter, query *dns.Msg) {
|
||||
startTime := time.Now()
|
||||
logger := log.WithFields(log.Fields{
|
||||
"request_id": resutil.GenerateRequestID(),
|
||||
"dns_id": fmt.Sprintf("%04x", query.Id),
|
||||
})
|
||||
|
||||
resp := f.handleDNSQuery(logger, w, query)
|
||||
resp := f.handleDNSQuery(w, query)
|
||||
if resp == nil {
|
||||
return
|
||||
}
|
||||
|
||||
if err := w.WriteMsg(resp); err != nil {
|
||||
logger.Errorf("failed to write DNS response: %v", err)
|
||||
return
|
||||
log.Errorf("failed to write DNS response: %v", err)
|
||||
}
|
||||
|
||||
logger.Tracef("response: domain=%s rcode=%s answers=%s took=%s",
|
||||
query.Question[0].Name, dns.RcodeToString[resp.Rcode], resutil.FormatAnswers(resp.Answer), time.Since(startTime))
|
||||
}
|
||||
|
||||
func (f *DNSForwarder) updateInternalState(ips []netip.Addr, mostSpecificResId route.ResID, matchingEntries []*ForwarderEntry) {
|
||||
@@ -325,64 +315,140 @@ func (f *DNSForwarder) updateFirewall(matchingEntries []*ForwarderEntry, prefixe
|
||||
}
|
||||
}
|
||||
|
||||
// setResponseCodeForNotFound determines and sets the appropriate response code when IsNotFound is true
|
||||
// It distinguishes between NXDOMAIN (domain doesn't exist) and NODATA (domain exists but no records of requested type)
|
||||
//
|
||||
// LIMITATION: This function only checks A and AAAA record types to determine domain existence.
|
||||
// If a domain has only other record types (MX, TXT, CNAME, etc.) but no A/AAAA records,
|
||||
// it may incorrectly return NXDOMAIN instead of NODATA. This is acceptable since the forwarder
|
||||
// only handles A/AAAA queries and returns NOTIMP for other types.
|
||||
func (f *DNSForwarder) setResponseCodeForNotFound(ctx context.Context, resp *dns.Msg, domain string, originalQtype uint16) {
|
||||
// Try querying for a different record type to see if the domain exists
|
||||
// If the original query was for AAAA, try A. If it was for A, try AAAA.
|
||||
// This helps distinguish between NXDOMAIN and NODATA.
|
||||
var alternativeNetwork string
|
||||
switch originalQtype {
|
||||
case dns.TypeAAAA:
|
||||
alternativeNetwork = "ip4"
|
||||
case dns.TypeA:
|
||||
alternativeNetwork = "ip6"
|
||||
default:
|
||||
resp.Rcode = dns.RcodeNameError
|
||||
return
|
||||
}
|
||||
|
||||
if _, err := f.resolver.LookupNetIP(ctx, alternativeNetwork, domain); err != nil {
|
||||
var dnsErr *net.DNSError
|
||||
if errors.As(err, &dnsErr) && dnsErr.IsNotFound {
|
||||
// Alternative query also returned not found - domain truly doesn't exist
|
||||
resp.Rcode = dns.RcodeNameError
|
||||
return
|
||||
}
|
||||
// Some other error (timeout, server failure, etc.) - can't determine, assume domain exists
|
||||
resp.Rcode = dns.RcodeSuccess
|
||||
return
|
||||
}
|
||||
|
||||
// Alternative query succeeded - domain exists but has no records of this type
|
||||
resp.Rcode = dns.RcodeSuccess
|
||||
}
|
||||
|
||||
// handleDNSError processes DNS lookup errors and sends an appropriate error response.
|
||||
func (f *DNSForwarder) handleDNSError(
|
||||
ctx context.Context,
|
||||
logger *log.Entry,
|
||||
w dns.ResponseWriter,
|
||||
question dns.Question,
|
||||
resp *dns.Msg,
|
||||
domain string,
|
||||
result resutil.LookupResult,
|
||||
err error,
|
||||
) {
|
||||
// Default to SERVFAIL; override below when appropriate.
|
||||
resp.Rcode = dns.RcodeServerFailure
|
||||
|
||||
qType := question.Qtype
|
||||
qTypeName := dns.TypeToString[qType]
|
||||
|
||||
resp.Rcode = result.Rcode
|
||||
|
||||
// NotFound: cache negative result and respond
|
||||
if result.Rcode == dns.RcodeNameError || result.Rcode == dns.RcodeSuccess {
|
||||
f.cache.set(domain, question.Qtype, nil)
|
||||
// Prefer typed DNS errors; fall back to generic logging otherwise.
|
||||
var dnsErr *net.DNSError
|
||||
if !errors.As(err, &dnsErr) {
|
||||
log.Warnf(errResolveFailed, domain, err)
|
||||
if writeErr := w.WriteMsg(resp); writeErr != nil {
|
||||
logger.Errorf("failed to write failure DNS response: %v", writeErr)
|
||||
log.Errorf("failed to write failure DNS response: %v", writeErr)
|
||||
}
|
||||
return
|
||||
}
|
||||
|
||||
// NotFound: set NXDOMAIN / appropriate code via helper.
|
||||
if dnsErr.IsNotFound {
|
||||
f.setResponseCodeForNotFound(ctx, resp, domain, qType)
|
||||
if writeErr := w.WriteMsg(resp); writeErr != nil {
|
||||
log.Errorf("failed to write failure DNS response: %v", writeErr)
|
||||
}
|
||||
f.cache.set(domain, question.Qtype, nil)
|
||||
return
|
||||
}
|
||||
|
||||
// Upstream failed but we might have a cached answer—serve it if present.
|
||||
if ips, ok := f.cache.get(domain, qType); ok {
|
||||
if len(ips) > 0 {
|
||||
logger.Debugf("serving cached DNS response after upstream failure: domain=%s type=%s", domain, qTypeName)
|
||||
resp.Answer = append(resp.Answer, resutil.IPsToRRs(domain, ips, f.ttl)...)
|
||||
log.Debugf("serving cached DNS response after upstream failure: domain=%s type=%s", domain, qTypeName)
|
||||
f.addIPsToResponse(resp, domain, ips)
|
||||
resp.Rcode = dns.RcodeSuccess
|
||||
if writeErr := w.WriteMsg(resp); writeErr != nil {
|
||||
logger.Errorf("failed to write cached DNS response: %v", writeErr)
|
||||
log.Errorf("failed to write cached DNS response: %v", writeErr)
|
||||
}
|
||||
return
|
||||
}
|
||||
|
||||
// Cached negative result - re-verify NXDOMAIN vs NODATA
|
||||
verifyResult := resutil.LookupIP(ctx, f.resolver, resutil.NetworkForQtype(qType), domain, qType)
|
||||
if verifyResult.Rcode == dns.RcodeNameError || verifyResult.Rcode == dns.RcodeSuccess {
|
||||
resp.Rcode = verifyResult.Rcode
|
||||
} else { // send NXDOMAIN / appropriate code if cache is empty
|
||||
f.setResponseCodeForNotFound(ctx, resp, domain, qType)
|
||||
if writeErr := w.WriteMsg(resp); writeErr != nil {
|
||||
logger.Errorf("failed to write failure DNS response: %v", writeErr)
|
||||
log.Errorf("failed to write failure DNS response: %v", writeErr)
|
||||
}
|
||||
return
|
||||
}
|
||||
return
|
||||
}
|
||||
|
||||
// No cache or verification failed. Log with or without the server field for more context.
|
||||
var dnsErr *net.DNSError
|
||||
if errors.As(result.Err, &dnsErr) && dnsErr.Server != "" {
|
||||
logger.Warnf("failed to resolve: type=%s domain=%s server=%s: %v", qTypeName, domain, dnsErr.Server, result.Err)
|
||||
// No cache. Log with or without the server field for more context.
|
||||
if dnsErr.Server != "" {
|
||||
log.Warnf("failed to resolve: type=%s domain=%s server=%s: %v", qTypeName, domain, dnsErr.Server, err)
|
||||
} else {
|
||||
logger.Warnf(errResolveFailed, domain, result.Err)
|
||||
log.Warnf(errResolveFailed, domain, err)
|
||||
}
|
||||
|
||||
// Write final failure response.
|
||||
if writeErr := w.WriteMsg(resp); writeErr != nil {
|
||||
logger.Errorf("failed to write failure DNS response: %v", writeErr)
|
||||
log.Errorf("failed to write failure DNS response: %v", writeErr)
|
||||
}
|
||||
}
|
||||
|
||||
// addIPsToResponse adds IP addresses to the DNS response as appropriate A or AAAA records
|
||||
func (f *DNSForwarder) addIPsToResponse(resp *dns.Msg, domain string, ips []netip.Addr) {
|
||||
for _, ip := range ips {
|
||||
var respRecord dns.RR
|
||||
if ip.Is6() {
|
||||
log.Tracef("resolved domain=%s to IPv6=%s", domain, ip)
|
||||
rr := dns.AAAA{
|
||||
AAAA: ip.AsSlice(),
|
||||
Hdr: dns.RR_Header{
|
||||
Name: domain,
|
||||
Rrtype: dns.TypeAAAA,
|
||||
Class: dns.ClassINET,
|
||||
Ttl: f.ttl,
|
||||
},
|
||||
}
|
||||
respRecord = &rr
|
||||
} else {
|
||||
log.Tracef("resolved domain=%s to IPv4=%s", domain, ip)
|
||||
rr := dns.A{
|
||||
A: ip.AsSlice(),
|
||||
Hdr: dns.RR_Header{
|
||||
Name: domain,
|
||||
Rrtype: dns.TypeA,
|
||||
Class: dns.ClassINET,
|
||||
Ttl: f.ttl,
|
||||
},
|
||||
}
|
||||
respRecord = &rr
|
||||
}
|
||||
resp.Answer = append(resp.Answer, respRecord)
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
@@ -10,7 +10,6 @@ import (
|
||||
"time"
|
||||
|
||||
"github.com/miekg/dns"
|
||||
log "github.com/sirupsen/logrus"
|
||||
"github.com/stretchr/testify/assert"
|
||||
"github.com/stretchr/testify/mock"
|
||||
"github.com/stretchr/testify/require"
|
||||
@@ -318,7 +317,7 @@ func TestDNSForwarder_UnauthorizedDomainAccess(t *testing.T) {
|
||||
query.SetQuestion(dns.Fqdn(tt.queryDomain), dns.TypeA)
|
||||
|
||||
mockWriter := &test.MockResponseWriter{}
|
||||
resp := forwarder.handleDNSQuery(log.NewEntry(log.StandardLogger()), mockWriter, query)
|
||||
resp := forwarder.handleDNSQuery(mockWriter, query)
|
||||
|
||||
if tt.shouldResolve {
|
||||
require.NotNil(t, resp, "Expected response for authorized domain")
|
||||
@@ -466,7 +465,7 @@ func TestDNSForwarder_FirewallSetUpdates(t *testing.T) {
|
||||
dnsQuery.SetQuestion(dns.Fqdn(tt.query), dns.TypeA)
|
||||
|
||||
mockWriter := &test.MockResponseWriter{}
|
||||
resp := forwarder.handleDNSQuery(log.NewEntry(log.StandardLogger()), mockWriter, dnsQuery)
|
||||
resp := forwarder.handleDNSQuery(mockWriter, dnsQuery)
|
||||
|
||||
// Verify response
|
||||
if tt.shouldResolve {
|
||||
@@ -528,7 +527,7 @@ func TestDNSForwarder_MultipleIPsInSingleUpdate(t *testing.T) {
|
||||
query.SetQuestion("example.com.", dns.TypeA)
|
||||
|
||||
mockWriter := &test.MockResponseWriter{}
|
||||
resp := forwarder.handleDNSQuery(log.NewEntry(log.StandardLogger()), mockWriter, query)
|
||||
resp := forwarder.handleDNSQuery(mockWriter, query)
|
||||
|
||||
// Verify response contains all IPs
|
||||
require.NotNil(t, resp)
|
||||
@@ -605,7 +604,7 @@ func TestDNSForwarder_ResponseCodes(t *testing.T) {
|
||||
},
|
||||
}
|
||||
|
||||
_ = forwarder.handleDNSQuery(log.NewEntry(log.StandardLogger()), mockWriter, query)
|
||||
_ = forwarder.handleDNSQuery(mockWriter, query)
|
||||
|
||||
// Check the response written to the writer
|
||||
require.NotNil(t, writtenResp, "Expected response to be written")
|
||||
@@ -675,7 +674,7 @@ func TestDNSForwarder_ServeFromCacheOnUpstreamFailure(t *testing.T) {
|
||||
q1 := &dns.Msg{}
|
||||
q1.SetQuestion(dns.Fqdn("example.com"), dns.TypeA)
|
||||
w1 := &test.MockResponseWriter{}
|
||||
resp1 := forwarder.handleDNSQuery(log.NewEntry(log.StandardLogger()), w1, q1)
|
||||
resp1 := forwarder.handleDNSQuery(w1, q1)
|
||||
require.NotNil(t, resp1)
|
||||
require.Equal(t, dns.RcodeSuccess, resp1.Rcode)
|
||||
require.Len(t, resp1.Answer, 1)
|
||||
@@ -685,7 +684,7 @@ func TestDNSForwarder_ServeFromCacheOnUpstreamFailure(t *testing.T) {
|
||||
q2.SetQuestion(dns.Fqdn("example.com"), dns.TypeA)
|
||||
var writtenResp *dns.Msg
|
||||
w2 := &test.MockResponseWriter{WriteMsgFunc: func(m *dns.Msg) error { writtenResp = m; return nil }}
|
||||
_ = forwarder.handleDNSQuery(log.NewEntry(log.StandardLogger()), w2, q2)
|
||||
_ = forwarder.handleDNSQuery(w2, q2)
|
||||
|
||||
require.NotNil(t, writtenResp, "expected response to be written")
|
||||
require.Equal(t, dns.RcodeSuccess, writtenResp.Rcode)
|
||||
@@ -715,7 +714,7 @@ func TestDNSForwarder_CacheNormalizationCasingAndDot(t *testing.T) {
|
||||
q1 := &dns.Msg{}
|
||||
q1.SetQuestion(mixedQuery+".", dns.TypeA)
|
||||
w1 := &test.MockResponseWriter{}
|
||||
resp1 := forwarder.handleDNSQuery(log.NewEntry(log.StandardLogger()), w1, q1)
|
||||
resp1 := forwarder.handleDNSQuery(w1, q1)
|
||||
require.NotNil(t, resp1)
|
||||
require.Equal(t, dns.RcodeSuccess, resp1.Rcode)
|
||||
require.Len(t, resp1.Answer, 1)
|
||||
@@ -729,7 +728,7 @@ func TestDNSForwarder_CacheNormalizationCasingAndDot(t *testing.T) {
|
||||
q2.SetQuestion("EXAMPLE.COM", dns.TypeA)
|
||||
var writtenResp *dns.Msg
|
||||
w2 := &test.MockResponseWriter{WriteMsgFunc: func(m *dns.Msg) error { writtenResp = m; return nil }}
|
||||
_ = forwarder.handleDNSQuery(log.NewEntry(log.StandardLogger()), w2, q2)
|
||||
_ = forwarder.handleDNSQuery(w2, q2)
|
||||
|
||||
require.NotNil(t, writtenResp)
|
||||
require.Equal(t, dns.RcodeSuccess, writtenResp.Rcode)
|
||||
@@ -784,7 +783,7 @@ func TestDNSForwarder_MultipleOverlappingPatterns(t *testing.T) {
|
||||
query.SetQuestion("smtp.mail.example.com.", dns.TypeA)
|
||||
|
||||
mockWriter := &test.MockResponseWriter{}
|
||||
resp := forwarder.handleDNSQuery(log.NewEntry(log.StandardLogger()), mockWriter, query)
|
||||
resp := forwarder.handleDNSQuery(mockWriter, query)
|
||||
|
||||
require.NotNil(t, resp)
|
||||
assert.Equal(t, dns.RcodeSuccess, resp.Rcode)
|
||||
@@ -905,7 +904,7 @@ func TestDNSForwarder_NodataVsNxdomain(t *testing.T) {
|
||||
},
|
||||
}
|
||||
|
||||
resp := forwarder.handleDNSQuery(log.NewEntry(log.StandardLogger()), mockWriter, query)
|
||||
resp := forwarder.handleDNSQuery(mockWriter, query)
|
||||
|
||||
// If a response was returned, it means it should be written (happens in wrapper functions)
|
||||
if resp != nil && writtenResp == nil {
|
||||
@@ -938,7 +937,7 @@ func TestDNSForwarder_EmptyQuery(t *testing.T) {
|
||||
return nil
|
||||
},
|
||||
}
|
||||
resp := forwarder.handleDNSQuery(log.NewEntry(log.StandardLogger()), mockWriter, query)
|
||||
resp := forwarder.handleDNSQuery(mockWriter, query)
|
||||
|
||||
assert.Nil(t, resp, "Should return nil for empty query")
|
||||
assert.False(t, writeCalled, "Should not write response for empty query")
|
||||
|
||||
@@ -42,13 +42,14 @@ import (
|
||||
"github.com/netbirdio/netbird/client/internal/peer/guard"
|
||||
icemaker "github.com/netbirdio/netbird/client/internal/peer/ice"
|
||||
"github.com/netbirdio/netbird/client/internal/peerstore"
|
||||
"github.com/netbirdio/netbird/client/internal/profilemanager"
|
||||
"github.com/netbirdio/netbird/client/internal/relay"
|
||||
"github.com/netbirdio/netbird/client/internal/rosenpass"
|
||||
"github.com/netbirdio/netbird/client/internal/routemanager"
|
||||
"github.com/netbirdio/netbird/client/internal/routemanager/systemops"
|
||||
"github.com/netbirdio/netbird/client/internal/statemanager"
|
||||
"github.com/netbirdio/netbird/client/internal/updatemanager"
|
||||
cProto "github.com/netbirdio/netbird/client/proto"
|
||||
sshconfig "github.com/netbirdio/netbird/client/ssh/config"
|
||||
"github.com/netbirdio/netbird/shared/management/domain"
|
||||
semaphoregroup "github.com/netbirdio/netbird/util/semaphore-group"
|
||||
|
||||
@@ -72,7 +73,6 @@ const (
|
||||
PeerConnectionTimeoutMax = 45000 // ms
|
||||
PeerConnectionTimeoutMin = 30000 // ms
|
||||
connInitLimit = 200
|
||||
disableAutoUpdate = "disabled"
|
||||
)
|
||||
|
||||
var ErrResetConnection = fmt.Errorf("reset connection")
|
||||
@@ -201,9 +201,6 @@ type Engine struct {
|
||||
connSemaphore *semaphoregroup.SemaphoreGroup
|
||||
flowManager nftypes.FlowManager
|
||||
|
||||
// auto-update
|
||||
updateManager *updatemanager.Manager
|
||||
|
||||
// WireGuard interface monitor
|
||||
wgIfaceMonitor *WGIfaceMonitor
|
||||
|
||||
@@ -224,7 +221,17 @@ type localIpUpdater interface {
|
||||
}
|
||||
|
||||
// NewEngine creates a new Connection Engine with probes attached
|
||||
func NewEngine(clientCtx context.Context, clientCancel context.CancelFunc, signalClient signal.Client, mgmClient mgm.Client, relayManager *relayClient.Manager, config *EngineConfig, mobileDep MobileDependency, statusRecorder *peer.Status, checks []*mgmProto.Checks, stateManager *statemanager.Manager) *Engine {
|
||||
func NewEngine(
|
||||
clientCtx context.Context,
|
||||
clientCancel context.CancelFunc,
|
||||
signalClient signal.Client,
|
||||
mgmClient mgm.Client,
|
||||
relayManager *relayClient.Manager,
|
||||
config *EngineConfig,
|
||||
mobileDep MobileDependency,
|
||||
statusRecorder *peer.Status,
|
||||
checks []*mgmProto.Checks,
|
||||
) *Engine {
|
||||
engine := &Engine{
|
||||
clientCtx: clientCtx,
|
||||
clientCancel: clientCancel,
|
||||
@@ -240,12 +247,28 @@ func NewEngine(clientCtx context.Context, clientCancel context.CancelFunc, signa
|
||||
TURNs: []*stun.URI{},
|
||||
networkSerial: 0,
|
||||
statusRecorder: statusRecorder,
|
||||
stateManager: stateManager,
|
||||
checks: checks,
|
||||
connSemaphore: semaphoregroup.NewSemaphoreGroup(connInitLimit),
|
||||
probeStunTurn: relay.NewStunTurnProbe(relay.DefaultCacheTTL),
|
||||
}
|
||||
|
||||
sm := profilemanager.NewServiceManager("")
|
||||
|
||||
path := sm.GetStatePath()
|
||||
if runtime.GOOS == "ios" || runtime.GOOS == "android" {
|
||||
if !fileExists(mobileDep.StateFilePath) {
|
||||
err := createFile(mobileDep.StateFilePath)
|
||||
if err != nil {
|
||||
log.Errorf("failed to create state file: %v", err)
|
||||
// we are not exiting as we can run without the state manager
|
||||
}
|
||||
}
|
||||
|
||||
path = mobileDep.StateFilePath
|
||||
}
|
||||
engine.stateManager = statemanager.New(path)
|
||||
engine.stateManager.RegisterState(&sshconfig.ShutdownState{})
|
||||
|
||||
log.Infof("I am: %s", config.WgPrivateKey.PublicKey().String())
|
||||
return engine
|
||||
}
|
||||
@@ -285,10 +308,6 @@ func (e *Engine) Stop() error {
|
||||
e.srWatcher.Close()
|
||||
}
|
||||
|
||||
if e.updateManager != nil {
|
||||
e.updateManager.Stop()
|
||||
}
|
||||
|
||||
log.Info("cleaning up status recorder states")
|
||||
e.statusRecorder.ReplaceOfflinePeers([]peer.State{})
|
||||
e.statusRecorder.UpdateDNSStates([]peer.NSGroupState{})
|
||||
@@ -522,13 +541,6 @@ func (e *Engine) Start(netbirdConfig *mgmProto.NetbirdConfig, mgmtURL *url.URL)
|
||||
return nil
|
||||
}
|
||||
|
||||
func (e *Engine) InitialUpdateHandling(autoUpdateSettings *mgmProto.AutoUpdateSettings) {
|
||||
e.syncMsgMux.Lock()
|
||||
defer e.syncMsgMux.Unlock()
|
||||
|
||||
e.handleAutoUpdateVersion(autoUpdateSettings, true)
|
||||
}
|
||||
|
||||
func (e *Engine) createFirewall() error {
|
||||
if e.config.DisableFirewall {
|
||||
log.Infof("firewall is disabled")
|
||||
@@ -737,41 +749,6 @@ func (e *Engine) PopulateNetbirdConfig(netbirdConfig *mgmProto.NetbirdConfig, mg
|
||||
return nil
|
||||
}
|
||||
|
||||
func (e *Engine) handleAutoUpdateVersion(autoUpdateSettings *mgmProto.AutoUpdateSettings, initialCheck bool) {
|
||||
if autoUpdateSettings == nil {
|
||||
return
|
||||
}
|
||||
|
||||
disabled := autoUpdateSettings.Version == disableAutoUpdate
|
||||
|
||||
// Stop and cleanup if disabled
|
||||
if e.updateManager != nil && disabled {
|
||||
log.Infof("auto-update is disabled, stopping update manager")
|
||||
e.updateManager.Stop()
|
||||
e.updateManager = nil
|
||||
return
|
||||
}
|
||||
|
||||
// Skip check unless AlwaysUpdate is enabled or this is the initial check at startup
|
||||
if !autoUpdateSettings.AlwaysUpdate && !initialCheck {
|
||||
log.Debugf("skipping auto-update check, AlwaysUpdate is false and this is not the initial check")
|
||||
return
|
||||
}
|
||||
|
||||
// Start manager if needed
|
||||
if e.updateManager == nil {
|
||||
log.Infof("starting auto-update manager")
|
||||
updateManager, err := updatemanager.NewManager(e.statusRecorder, e.stateManager)
|
||||
if err != nil {
|
||||
return
|
||||
}
|
||||
e.updateManager = updateManager
|
||||
e.updateManager.Start(e.ctx)
|
||||
}
|
||||
log.Infof("handling auto-update version: %s", autoUpdateSettings.Version)
|
||||
e.updateManager.SetVersion(autoUpdateSettings.Version)
|
||||
}
|
||||
|
||||
func (e *Engine) handleSync(update *mgmProto.SyncResponse) error {
|
||||
e.syncMsgMux.Lock()
|
||||
defer e.syncMsgMux.Unlock()
|
||||
@@ -781,10 +758,6 @@ func (e *Engine) handleSync(update *mgmProto.SyncResponse) error {
|
||||
return e.ctx.Err()
|
||||
}
|
||||
|
||||
if update.NetworkMap != nil && update.NetworkMap.PeerConfig != nil {
|
||||
e.handleAutoUpdateVersion(update.NetworkMap.PeerConfig.AutoUpdate, false)
|
||||
}
|
||||
|
||||
if update.GetNetbirdConfig() != nil {
|
||||
wCfg := update.GetNetbirdConfig()
|
||||
err := e.updateTURNs(wCfg.GetTurns())
|
||||
@@ -824,7 +797,7 @@ func (e *Engine) handleSync(update *mgmProto.SyncResponse) error {
|
||||
}
|
||||
|
||||
nm := update.GetNetworkMap()
|
||||
if nm == nil {
|
||||
if nm == nil || update.SkipNetworkMapUpdate {
|
||||
return nil
|
||||
}
|
||||
|
||||
@@ -990,7 +963,7 @@ func (e *Engine) receiveManagementEvents() {
|
||||
e.config.DisableSSHAuth,
|
||||
)
|
||||
|
||||
err = e.mgmClient.Sync(e.ctx, info, e.handleSync)
|
||||
err = e.mgmClient.Sync(e.ctx, info, e.networkSerial, e.handleSync)
|
||||
if err != nil {
|
||||
// happens if management is unavailable for a long time.
|
||||
// We want to cancel the operation of the whole client
|
||||
@@ -1121,15 +1094,6 @@ func (e *Engine) updateNetworkMap(networkMap *mgmProto.NetworkMap) error {
|
||||
|
||||
e.updateOfflinePeers(networkMap.GetOfflinePeers())
|
||||
|
||||
// Filter out own peer from the remote peers list
|
||||
localPubKey := e.config.WgPrivateKey.PublicKey().String()
|
||||
remotePeers := make([]*mgmProto.RemotePeerConfig, 0, len(networkMap.GetRemotePeers()))
|
||||
for _, p := range networkMap.GetRemotePeers() {
|
||||
if p.GetWgPubKey() != localPubKey {
|
||||
remotePeers = append(remotePeers, p)
|
||||
}
|
||||
}
|
||||
|
||||
// cleanup request, most likely our peer has been deleted
|
||||
if networkMap.GetRemotePeersIsEmpty() {
|
||||
err := e.removeAllPeers()
|
||||
@@ -1138,34 +1102,32 @@ func (e *Engine) updateNetworkMap(networkMap *mgmProto.NetworkMap) error {
|
||||
return err
|
||||
}
|
||||
} else {
|
||||
err := e.removePeers(remotePeers)
|
||||
err := e.removePeers(networkMap.GetRemotePeers())
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
|
||||
err = e.modifyPeers(remotePeers)
|
||||
err = e.modifyPeers(networkMap.GetRemotePeers())
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
|
||||
err = e.addNewPeers(remotePeers)
|
||||
err = e.addNewPeers(networkMap.GetRemotePeers())
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
|
||||
e.statusRecorder.FinishPeerListModifications()
|
||||
|
||||
e.updatePeerSSHHostKeys(remotePeers)
|
||||
e.updatePeerSSHHostKeys(networkMap.GetRemotePeers())
|
||||
|
||||
if err := e.updateSSHClientConfig(remotePeers); err != nil {
|
||||
if err := e.updateSSHClientConfig(networkMap.GetRemotePeers()); err != nil {
|
||||
log.Warnf("failed to update SSH client config: %v", err)
|
||||
}
|
||||
|
||||
e.updateSSHServerAuth(networkMap.GetSshAuth())
|
||||
}
|
||||
|
||||
// must set the exclude list after the peers are added. Without it the manager can not figure out the peers parameters from the store
|
||||
excludedLazyPeers := e.toExcludedLazyPeers(forwardingRules, remotePeers)
|
||||
excludedLazyPeers := e.toExcludedLazyPeers(forwardingRules, networkMap.GetRemotePeers())
|
||||
e.connMgr.SetExcludeList(e.ctx, excludedLazyPeers)
|
||||
|
||||
e.networkSerial = serial
|
||||
@@ -1251,16 +1213,11 @@ func toDNSConfig(protoDNSConfig *mgmProto.DNSConfig, network netip.Prefix) nbdns
|
||||
ForwarderPort: forwarderPort,
|
||||
}
|
||||
|
||||
protoZones := protoDNSConfig.GetCustomZones()
|
||||
// Treat single zone as authoritative for backward compatibility with old servers
|
||||
// that only send the peer FQDN zone without setting field 4.
|
||||
singleZoneCompat := len(protoZones) == 1
|
||||
|
||||
for _, zone := range protoZones {
|
||||
for _, zone := range protoDNSConfig.GetCustomZones() {
|
||||
dnsZone := nbdns.CustomZone{
|
||||
Domain: zone.GetDomain(),
|
||||
SearchDomainDisabled: zone.GetSearchDomainDisabled(),
|
||||
NonAuthoritative: zone.GetNonAuthoritative() && !singleZoneCompat,
|
||||
SkipPTRProcess: zone.GetSkipPTRProcess(),
|
||||
}
|
||||
for _, record := range zone.Records {
|
||||
dnsRecord := nbdns.SimpleRecord{
|
||||
@@ -1748,26 +1705,22 @@ func (e *Engine) RunHealthProbes(waitForResult bool) bool {
|
||||
}
|
||||
|
||||
e.syncMsgMux.Unlock()
|
||||
|
||||
// Skip STUN/TURN probing for JS/WASM as it's not available
|
||||
relayHealthy := true
|
||||
if runtime.GOOS != "js" {
|
||||
var results []relay.ProbeResult
|
||||
if waitForResult {
|
||||
results = e.probeStunTurn.ProbeAllWaitResult(e.ctx, stuns, turns)
|
||||
} else {
|
||||
results = e.probeStunTurn.ProbeAll(e.ctx, stuns, turns)
|
||||
}
|
||||
e.statusRecorder.UpdateRelayStates(results)
|
||||
|
||||
for _, res := range results {
|
||||
if res.Err != nil {
|
||||
relayHealthy = false
|
||||
break
|
||||
}
|
||||
}
|
||||
log.Debugf("relay health check: healthy=%t", relayHealthy)
|
||||
var results []relay.ProbeResult
|
||||
if waitForResult {
|
||||
results = e.probeStunTurn.ProbeAllWaitResult(e.ctx, stuns, turns)
|
||||
} else {
|
||||
results = e.probeStunTurn.ProbeAll(e.ctx, stuns, turns)
|
||||
}
|
||||
e.statusRecorder.UpdateRelayStates(results)
|
||||
|
||||
relayHealthy := true
|
||||
for _, res := range results {
|
||||
if res.Err != nil {
|
||||
relayHealthy = false
|
||||
break
|
||||
}
|
||||
}
|
||||
log.Debugf("relay health check: healthy=%t", relayHealthy)
|
||||
|
||||
allHealthy := signalHealthy && managementHealthy && relayHealthy
|
||||
log.Debugf("all health checks completed: healthy=%t", allHealthy)
|
||||
|
||||
@@ -11,18 +11,15 @@ import (
|
||||
|
||||
firewallManager "github.com/netbirdio/netbird/client/firewall/manager"
|
||||
nftypes "github.com/netbirdio/netbird/client/internal/netflow/types"
|
||||
sshauth "github.com/netbirdio/netbird/client/ssh/auth"
|
||||
sshconfig "github.com/netbirdio/netbird/client/ssh/config"
|
||||
sshserver "github.com/netbirdio/netbird/client/ssh/server"
|
||||
mgmProto "github.com/netbirdio/netbird/shared/management/proto"
|
||||
sshuserhash "github.com/netbirdio/netbird/shared/sshauth"
|
||||
)
|
||||
|
||||
type sshServer interface {
|
||||
Start(ctx context.Context, addr netip.AddrPort) error
|
||||
Stop() error
|
||||
GetStatus() (bool, []sshserver.SessionInfo)
|
||||
UpdateSSHAuth(config *sshauth.Config)
|
||||
}
|
||||
|
||||
func (e *Engine) setupSSHPortRedirection() error {
|
||||
@@ -72,16 +69,9 @@ func (e *Engine) updateSSH(sshConf *mgmProto.SSHConfig) error {
|
||||
}
|
||||
|
||||
if protoJWT := sshConf.GetJwtConfig(); protoJWT != nil {
|
||||
audiences := protoJWT.GetAudiences()
|
||||
if len(audiences) == 0 && protoJWT.GetAudience() != "" {
|
||||
audiences = []string{protoJWT.GetAudience()}
|
||||
}
|
||||
|
||||
log.Debugf("starting SSH server with JWT authentication: audiences=%v", audiences)
|
||||
|
||||
jwtConfig := &sshserver.JWTConfig{
|
||||
Issuer: protoJWT.GetIssuer(),
|
||||
Audiences: audiences,
|
||||
Audience: protoJWT.GetAudience(),
|
||||
KeysLocation: protoJWT.GetKeysLocation(),
|
||||
MaxTokenAge: protoJWT.GetMaxTokenAge(),
|
||||
}
|
||||
@@ -363,38 +353,3 @@ func (e *Engine) GetSSHServerStatus() (enabled bool, sessions []sshserver.Sessio
|
||||
|
||||
return sshServer.GetStatus()
|
||||
}
|
||||
|
||||
// updateSSHServerAuth updates SSH fine-grained access control configuration on a running SSH server
|
||||
func (e *Engine) updateSSHServerAuth(sshAuth *mgmProto.SSHAuth) {
|
||||
if sshAuth == nil {
|
||||
return
|
||||
}
|
||||
|
||||
if e.sshServer == nil {
|
||||
return
|
||||
}
|
||||
|
||||
protoUsers := sshAuth.GetAuthorizedUsers()
|
||||
authorizedUsers := make([]sshuserhash.UserIDHash, len(protoUsers))
|
||||
for i, hash := range protoUsers {
|
||||
if len(hash) != 16 {
|
||||
log.Warnf("invalid hash length %d, expected 16 - skipping SSH server auth update", len(hash))
|
||||
return
|
||||
}
|
||||
authorizedUsers[i] = sshuserhash.UserIDHash(hash)
|
||||
}
|
||||
|
||||
machineUsers := make(map[string][]uint32)
|
||||
for osUser, indexes := range sshAuth.GetMachineUsers() {
|
||||
machineUsers[osUser] = indexes.GetIndexes()
|
||||
}
|
||||
|
||||
// Update SSH server with new authorization configuration
|
||||
authConfig := &sshauth.Config{
|
||||
UserIDClaim: sshAuth.GetUserIDClaim(),
|
||||
AuthorizedUsers: authorizedUsers,
|
||||
MachineUsers: machineUsers,
|
||||
}
|
||||
|
||||
e.sshServer.UpdateSSHAuth(authConfig)
|
||||
}
|
||||
|
||||
79
client/internal/engine_sync_test.go
Normal file
79
client/internal/engine_sync_test.go
Normal file
@@ -0,0 +1,79 @@
|
||||
package internal
|
||||
|
||||
import (
|
||||
"context"
|
||||
"testing"
|
||||
|
||||
"golang.zx2c4.com/wireguard/wgctrl/wgtypes"
|
||||
|
||||
"github.com/netbirdio/netbird/client/iface"
|
||||
"github.com/netbirdio/netbird/client/internal/peer"
|
||||
"github.com/netbirdio/netbird/shared/management/client"
|
||||
mgmtProto "github.com/netbirdio/netbird/shared/management/proto"
|
||||
)
|
||||
|
||||
// Ensures handleSync exits early when SkipNetworkMapUpdate is true
|
||||
func TestEngine_HandleSync_SkipNetworkMapUpdate(t *testing.T) {
|
||||
key, err := wgtypes.GeneratePrivateKey()
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
|
||||
ctx, cancel := context.WithCancel(context.Background())
|
||||
defer cancel()
|
||||
|
||||
engine := NewEngine(ctx, cancel, nil, &client.MockClient{}, nil, &EngineConfig{
|
||||
WgIfaceName: "utun199",
|
||||
WgAddr: "100.70.0.1/24",
|
||||
WgPrivateKey: key,
|
||||
WgPort: 33100,
|
||||
MTU: iface.DefaultMTU,
|
||||
}, MobileDependency{}, peer.NewRecorder("https://mgm"), nil)
|
||||
engine.ctx = ctx
|
||||
|
||||
// Precondition
|
||||
if engine.networkSerial != 0 {
|
||||
t.Fatalf("unexpected initial serial: %d", engine.networkSerial)
|
||||
}
|
||||
|
||||
resp := &mgmtProto.SyncResponse{
|
||||
NetworkMap: &mgmtProto.NetworkMap{Serial: 42},
|
||||
SkipNetworkMapUpdate: true,
|
||||
}
|
||||
|
||||
if err := engine.handleSync(resp); err != nil {
|
||||
t.Fatalf("handleSync returned error: %v", err)
|
||||
}
|
||||
|
||||
if engine.networkSerial != 0 {
|
||||
t.Fatalf("networkSerial changed despite SkipNetworkMapUpdate; got %d, want 0", engine.networkSerial)
|
||||
}
|
||||
}
|
||||
|
||||
// Ensures handleSync exits early when NetworkMap is nil
|
||||
func TestEngine_HandleSync_NilNetworkMap(t *testing.T) {
|
||||
key, err := wgtypes.GeneratePrivateKey()
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
|
||||
ctx, cancel := context.WithCancel(context.Background())
|
||||
defer cancel()
|
||||
|
||||
engine := NewEngine(ctx, cancel, nil, &client.MockClient{}, nil, &EngineConfig{
|
||||
WgIfaceName: "utun198",
|
||||
WgAddr: "100.70.0.2/24",
|
||||
WgPrivateKey: key,
|
||||
WgPort: 33101,
|
||||
MTU: iface.DefaultMTU,
|
||||
}, MobileDependency{}, peer.NewRecorder("https://mgm"), nil)
|
||||
engine.ctx = ctx
|
||||
|
||||
resp := &mgmtProto.SyncResponse{NetworkMap: nil}
|
||||
|
||||
if err := engine.handleSync(resp); err != nil {
|
||||
t.Fatalf("handleSync returned error: %v", err)
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
@@ -253,7 +253,6 @@ func TestEngine_SSH(t *testing.T) {
|
||||
MobileDependency{},
|
||||
peer.NewRecorder("https://mgm"),
|
||||
nil,
|
||||
nil,
|
||||
)
|
||||
|
||||
engine.dnsServer = &dns.MockServer{
|
||||
@@ -415,13 +414,21 @@ func TestEngine_UpdateNetworkMap(t *testing.T) {
|
||||
defer cancel()
|
||||
|
||||
relayMgr := relayClient.NewManager(ctx, nil, key.PublicKey().String(), iface.DefaultMTU)
|
||||
engine := NewEngine(ctx, cancel, &signal.MockClient{}, &mgmt.MockClient{}, relayMgr, &EngineConfig{
|
||||
WgIfaceName: "utun102",
|
||||
WgAddr: "100.64.0.1/24",
|
||||
WgPrivateKey: key,
|
||||
WgPort: 33100,
|
||||
MTU: iface.DefaultMTU,
|
||||
}, MobileDependency{}, peer.NewRecorder("https://mgm"), nil, nil)
|
||||
engine := NewEngine(
|
||||
ctx, cancel,
|
||||
&signal.MockClient{},
|
||||
&mgmt.MockClient{},
|
||||
relayMgr,
|
||||
&EngineConfig{
|
||||
WgIfaceName: "utun102",
|
||||
WgAddr: "100.64.0.1/24",
|
||||
WgPrivateKey: key,
|
||||
WgPort: 33100,
|
||||
MTU: iface.DefaultMTU,
|
||||
},
|
||||
MobileDependency{},
|
||||
peer.NewRecorder("https://mgm"),
|
||||
nil)
|
||||
|
||||
wgIface := &MockWGIface{
|
||||
NameFunc: func() string { return "utun102" },
|
||||
@@ -624,7 +631,7 @@ func TestEngine_Sync(t *testing.T) {
|
||||
// feed updates to Engine via mocked Management client
|
||||
updates := make(chan *mgmtProto.SyncResponse)
|
||||
defer close(updates)
|
||||
syncFunc := func(ctx context.Context, info *system.Info, msgHandler func(msg *mgmtProto.SyncResponse) error) error {
|
||||
syncFunc := func(ctx context.Context, info *system.Info, networkSerial uint64, msgHandler func(msg *mgmtProto.SyncResponse) error) error {
|
||||
for msg := range updates {
|
||||
err := msgHandler(msg)
|
||||
if err != nil {
|
||||
@@ -640,7 +647,7 @@ func TestEngine_Sync(t *testing.T) {
|
||||
WgPrivateKey: key,
|
||||
WgPort: 33100,
|
||||
MTU: iface.DefaultMTU,
|
||||
}, MobileDependency{}, peer.NewRecorder("https://mgm"), nil, nil)
|
||||
}, MobileDependency{}, peer.NewRecorder("https://mgm"), nil)
|
||||
engine.ctx = ctx
|
||||
|
||||
engine.dnsServer = &dns.MockServer{
|
||||
@@ -805,7 +812,7 @@ func TestEngine_UpdateNetworkMapWithRoutes(t *testing.T) {
|
||||
WgPrivateKey: key,
|
||||
WgPort: 33100,
|
||||
MTU: iface.DefaultMTU,
|
||||
}, MobileDependency{}, peer.NewRecorder("https://mgm"), nil, nil)
|
||||
}, MobileDependency{}, peer.NewRecorder("https://mgm"), nil)
|
||||
engine.ctx = ctx
|
||||
newNet, err := stdnet.NewNet(context.Background(), nil)
|
||||
if err != nil {
|
||||
@@ -1007,7 +1014,7 @@ func TestEngine_UpdateNetworkMapWithDNSUpdate(t *testing.T) {
|
||||
WgPrivateKey: key,
|
||||
WgPort: 33100,
|
||||
MTU: iface.DefaultMTU,
|
||||
}, MobileDependency{}, peer.NewRecorder("https://mgm"), nil, nil)
|
||||
}, MobileDependency{}, peer.NewRecorder("https://mgm"), nil)
|
||||
engine.ctx = ctx
|
||||
|
||||
newNet, err := stdnet.NewNet(context.Background(), nil)
|
||||
@@ -1533,7 +1540,7 @@ func createEngine(ctx context.Context, cancel context.CancelFunc, setupKey strin
|
||||
}
|
||||
|
||||
relayMgr := relayClient.NewManager(ctx, nil, key.PublicKey().String(), iface.DefaultMTU)
|
||||
e, err := NewEngine(ctx, cancel, signalClient, mgmtClient, relayMgr, conf, MobileDependency{}, peer.NewRecorder("https://mgm"), nil, nil), nil
|
||||
e, err := NewEngine(ctx, cancel, signalClient, mgmtClient, relayMgr, conf, MobileDependency{}, peer.NewRecorder("https://mgm"), nil), nil
|
||||
e.ctx = ctx
|
||||
return e, err
|
||||
}
|
||||
@@ -1631,7 +1638,7 @@ func startManagement(t *testing.T, dataDir, testFile string) (*grpc.Server, stri
|
||||
if err != nil {
|
||||
return nil, "", err
|
||||
}
|
||||
mgmtServer, err := nbgrpc.NewServer(config, accountManager, settingsMockManager, secretsManager, nil, nil, &server.MockIntegratedValidator{}, networkMapController, nil)
|
||||
mgmtServer, err := nbgrpc.NewServer(config, accountManager, settingsMockManager, secretsManager, nil, nil, &server.MockIntegratedValidator{}, networkMapController)
|
||||
if err != nil {
|
||||
return nil, "", err
|
||||
}
|
||||
|
||||
@@ -1,4 +1,5 @@
|
||||
//go:build !windows
|
||||
// +build !windows
|
||||
|
||||
package internal
|
||||
|
||||
|
||||
@@ -110,6 +110,7 @@ func wakeUpListen(ctx context.Context) {
|
||||
}
|
||||
|
||||
if newHash == initialHash {
|
||||
log.Tracef("no wakeup detected")
|
||||
continue
|
||||
}
|
||||
|
||||
|
||||
@@ -148,15 +148,13 @@ func NewConn(config ConnConfig, services ServiceDependencies) (*Conn, error) {
|
||||
// It will try to establish a connection using ICE and in parallel with relay. The higher priority connection type will
|
||||
// be used.
|
||||
func (conn *Conn) Open(engineCtx context.Context) error {
|
||||
if err := conn.semaphore.Add(engineCtx); err != nil {
|
||||
return err
|
||||
}
|
||||
conn.semaphore.Add(engineCtx)
|
||||
|
||||
conn.mu.Lock()
|
||||
defer conn.mu.Unlock()
|
||||
|
||||
if conn.opened {
|
||||
conn.semaphore.Done()
|
||||
conn.semaphore.Done(engineCtx)
|
||||
return nil
|
||||
}
|
||||
|
||||
@@ -167,7 +165,6 @@ func (conn *Conn) Open(engineCtx context.Context) error {
|
||||
relayIsSupportedLocally := conn.workerRelay.RelayIsSupportedLocally()
|
||||
workerICE, err := NewWorkerICE(conn.ctx, conn.Log, conn.config, conn, conn.signaler, conn.iFaceDiscover, conn.statusRecorder, relayIsSupportedLocally)
|
||||
if err != nil {
|
||||
conn.semaphore.Done()
|
||||
return err
|
||||
}
|
||||
conn.workerICE = workerICE
|
||||
@@ -203,7 +200,7 @@ func (conn *Conn) Open(engineCtx context.Context) error {
|
||||
defer conn.wg.Done()
|
||||
|
||||
conn.waitInitialRandomSleepTime(conn.ctx)
|
||||
conn.semaphore.Done()
|
||||
conn.semaphore.Done(conn.ctx)
|
||||
|
||||
conn.guard.Start(conn.ctx, conn.onGuardEvent)
|
||||
}()
|
||||
@@ -669,17 +666,10 @@ func (conn *Conn) isConnectedOnAllWay() (connected bool) {
|
||||
}
|
||||
}()
|
||||
|
||||
// For JS platform: only relay connection is supported
|
||||
if runtime.GOOS == "js" {
|
||||
return conn.statusRelay.Get() == worker.StatusConnected
|
||||
}
|
||||
|
||||
// For non-JS platforms: check ICE connection status
|
||||
if conn.statusICE.Get() == worker.StatusDisconnected && !conn.workerICE.InProgress() {
|
||||
if runtime.GOOS != "js" && conn.statusICE.Get() == worker.StatusDisconnected && !conn.workerICE.InProgress() {
|
||||
return false
|
||||
}
|
||||
|
||||
// If relay is supported with peer, it must also be connected
|
||||
if conn.workerRelay.IsRelayConnectionSupportedWithPeer() {
|
||||
if conn.statusRelay.Get() == worker.StatusDisconnected {
|
||||
return false
|
||||
|
||||
@@ -20,7 +20,7 @@ type EndpointUpdater struct {
|
||||
wgConfig WgConfig
|
||||
initiator bool
|
||||
|
||||
// mu protects cancelFunc
|
||||
// mu protects updateWireGuardPeer and cancelFunc
|
||||
mu sync.Mutex
|
||||
cancelFunc func()
|
||||
updateWg sync.WaitGroup
|
||||
@@ -86,9 +86,11 @@ func (e *EndpointUpdater) scheduleDelayedUpdate(ctx context.Context, addr *net.U
|
||||
case <-ctx.Done():
|
||||
return
|
||||
case <-t.C:
|
||||
e.mu.Lock()
|
||||
if err := e.updateWireGuardPeer(addr, presharedKey); err != nil {
|
||||
e.log.Errorf("failed to update WireGuard peer, address: %s, error: %v", addr, err)
|
||||
}
|
||||
e.mu.Unlock()
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
@@ -14,7 +14,6 @@ import (
|
||||
"golang.org/x/exp/maps"
|
||||
"google.golang.org/grpc/codes"
|
||||
gstatus "google.golang.org/grpc/status"
|
||||
"google.golang.org/protobuf/types/known/durationpb"
|
||||
"google.golang.org/protobuf/types/known/timestamppb"
|
||||
|
||||
firewall "github.com/netbirdio/netbird/client/firewall/manager"
|
||||
@@ -159,7 +158,6 @@ type FullStatus struct {
|
||||
NSGroupStates []NSGroupState
|
||||
NumOfForwardingRules int
|
||||
LazyConnectionEnabled bool
|
||||
Events []*proto.SystemEvent
|
||||
}
|
||||
|
||||
type StatusChangeSubscription struct {
|
||||
@@ -983,7 +981,6 @@ func (d *Status) GetFullStatus() FullStatus {
|
||||
}
|
||||
|
||||
fullStatus.Peers = append(fullStatus.Peers, d.offlinePeers...)
|
||||
fullStatus.Events = d.GetEventHistory()
|
||||
return fullStatus
|
||||
}
|
||||
|
||||
@@ -1184,97 +1181,3 @@ type EventSubscription struct {
|
||||
func (s *EventSubscription) Events() <-chan *proto.SystemEvent {
|
||||
return s.events
|
||||
}
|
||||
|
||||
// ToProto converts FullStatus to proto.FullStatus.
|
||||
func (fs FullStatus) ToProto() *proto.FullStatus {
|
||||
pbFullStatus := proto.FullStatus{
|
||||
ManagementState: &proto.ManagementState{},
|
||||
SignalState: &proto.SignalState{},
|
||||
LocalPeerState: &proto.LocalPeerState{},
|
||||
Peers: []*proto.PeerState{},
|
||||
}
|
||||
|
||||
pbFullStatus.ManagementState.URL = fs.ManagementState.URL
|
||||
pbFullStatus.ManagementState.Connected = fs.ManagementState.Connected
|
||||
if err := fs.ManagementState.Error; err != nil {
|
||||
pbFullStatus.ManagementState.Error = err.Error()
|
||||
}
|
||||
|
||||
pbFullStatus.SignalState.URL = fs.SignalState.URL
|
||||
pbFullStatus.SignalState.Connected = fs.SignalState.Connected
|
||||
if err := fs.SignalState.Error; err != nil {
|
||||
pbFullStatus.SignalState.Error = err.Error()
|
||||
}
|
||||
|
||||
pbFullStatus.LocalPeerState.IP = fs.LocalPeerState.IP
|
||||
pbFullStatus.LocalPeerState.PubKey = fs.LocalPeerState.PubKey
|
||||
pbFullStatus.LocalPeerState.KernelInterface = fs.LocalPeerState.KernelInterface
|
||||
pbFullStatus.LocalPeerState.Fqdn = fs.LocalPeerState.FQDN
|
||||
pbFullStatus.LocalPeerState.RosenpassPermissive = fs.RosenpassState.Permissive
|
||||
pbFullStatus.LocalPeerState.RosenpassEnabled = fs.RosenpassState.Enabled
|
||||
pbFullStatus.NumberOfForwardingRules = int32(fs.NumOfForwardingRules)
|
||||
pbFullStatus.LazyConnectionEnabled = fs.LazyConnectionEnabled
|
||||
|
||||
pbFullStatus.LocalPeerState.Networks = maps.Keys(fs.LocalPeerState.Routes)
|
||||
|
||||
for _, peerState := range fs.Peers {
|
||||
networks := maps.Keys(peerState.GetRoutes())
|
||||
|
||||
pbPeerState := &proto.PeerState{
|
||||
IP: peerState.IP,
|
||||
PubKey: peerState.PubKey,
|
||||
ConnStatus: peerState.ConnStatus.String(),
|
||||
ConnStatusUpdate: timestamppb.New(peerState.ConnStatusUpdate),
|
||||
Relayed: peerState.Relayed,
|
||||
LocalIceCandidateType: peerState.LocalIceCandidateType,
|
||||
RemoteIceCandidateType: peerState.RemoteIceCandidateType,
|
||||
LocalIceCandidateEndpoint: peerState.LocalIceCandidateEndpoint,
|
||||
RemoteIceCandidateEndpoint: peerState.RemoteIceCandidateEndpoint,
|
||||
RelayAddress: peerState.RelayServerAddress,
|
||||
Fqdn: peerState.FQDN,
|
||||
LastWireguardHandshake: timestamppb.New(peerState.LastWireguardHandshake),
|
||||
BytesRx: peerState.BytesRx,
|
||||
BytesTx: peerState.BytesTx,
|
||||
RosenpassEnabled: peerState.RosenpassEnabled,
|
||||
Networks: networks,
|
||||
Latency: durationpb.New(peerState.Latency),
|
||||
SshHostKey: peerState.SSHHostKey,
|
||||
}
|
||||
pbFullStatus.Peers = append(pbFullStatus.Peers, pbPeerState)
|
||||
}
|
||||
|
||||
for _, relayState := range fs.Relays {
|
||||
pbRelayState := &proto.RelayState{
|
||||
URI: relayState.URI,
|
||||
Available: relayState.Err == nil,
|
||||
}
|
||||
if err := relayState.Err; err != nil {
|
||||
pbRelayState.Error = err.Error()
|
||||
}
|
||||
pbFullStatus.Relays = append(pbFullStatus.Relays, pbRelayState)
|
||||
}
|
||||
|
||||
for _, dnsState := range fs.NSGroupStates {
|
||||
var err string
|
||||
if dnsState.Error != nil {
|
||||
err = dnsState.Error.Error()
|
||||
}
|
||||
|
||||
var servers []string
|
||||
for _, server := range dnsState.Servers {
|
||||
servers = append(servers, server.String())
|
||||
}
|
||||
|
||||
pbDnsState := &proto.NSGroupState{
|
||||
Servers: servers,
|
||||
Domains: dnsState.Domains,
|
||||
Enabled: dnsState.Enabled,
|
||||
Error: err,
|
||||
}
|
||||
pbFullStatus.DnsServers = append(pbFullStatus.DnsServers, pbDnsState)
|
||||
}
|
||||
|
||||
pbFullStatus.Events = fs.Events
|
||||
|
||||
return &pbFullStatus
|
||||
}
|
||||
|
||||
@@ -3,11 +3,9 @@ package profilemanager
|
||||
import (
|
||||
"context"
|
||||
"crypto/tls"
|
||||
"encoding/json"
|
||||
"fmt"
|
||||
"net/url"
|
||||
"os"
|
||||
"os/user"
|
||||
"path/filepath"
|
||||
"reflect"
|
||||
"runtime"
|
||||
@@ -167,26 +165,19 @@ func getConfigDir() (string, error) {
|
||||
if ConfigDirOverride != "" {
|
||||
return ConfigDirOverride, nil
|
||||
}
|
||||
|
||||
base, err := baseConfigDir()
|
||||
configDir, err := os.UserConfigDir()
|
||||
if err != nil {
|
||||
return "", err
|
||||
}
|
||||
|
||||
configDir := filepath.Join(base, "netbird")
|
||||
if err := os.MkdirAll(configDir, 0o755); err != nil {
|
||||
return "", err
|
||||
}
|
||||
return configDir, nil
|
||||
}
|
||||
|
||||
func baseConfigDir() (string, error) {
|
||||
if runtime.GOOS == "darwin" {
|
||||
if u, err := user.Current(); err == nil && u.HomeDir != "" {
|
||||
return filepath.Join(u.HomeDir, "Library", "Application Support"), nil
|
||||
configDir = filepath.Join(configDir, "netbird")
|
||||
if _, err := os.Stat(configDir); os.IsNotExist(err) {
|
||||
if err := os.MkdirAll(configDir, 0755); err != nil {
|
||||
return "", err
|
||||
}
|
||||
}
|
||||
return os.UserConfigDir()
|
||||
|
||||
return configDir, nil
|
||||
}
|
||||
|
||||
func getConfigDirForUser(username string) (string, error) {
|
||||
@@ -685,7 +676,7 @@ func update(input ConfigInput) (*Config, error) {
|
||||
return config, nil
|
||||
}
|
||||
|
||||
// GetConfig read config file and return with Config and if it was created. Errors out if it does not exist
|
||||
// GetConfig read config file and return with Config. Errors out if it does not exist
|
||||
func GetConfig(configPath string) (*Config, error) {
|
||||
return readConfig(configPath, false)
|
||||
}
|
||||
@@ -821,85 +812,3 @@ func readConfig(configPath string, createIfMissing bool) (*Config, error) {
|
||||
func WriteOutConfig(path string, config *Config) error {
|
||||
return util.WriteJson(context.Background(), path, config)
|
||||
}
|
||||
|
||||
// DirectWriteOutConfig writes config directly without atomic temp file operations.
|
||||
// Use this on platforms where atomic writes are blocked (e.g., tvOS sandbox).
|
||||
func DirectWriteOutConfig(path string, config *Config) error {
|
||||
return util.DirectWriteJson(context.Background(), path, config)
|
||||
}
|
||||
|
||||
// DirectUpdateOrCreateConfig is like UpdateOrCreateConfig but uses direct (non-atomic) writes.
|
||||
// Use this on platforms where atomic writes are blocked (e.g., tvOS sandbox).
|
||||
func DirectUpdateOrCreateConfig(input ConfigInput) (*Config, error) {
|
||||
if !fileExists(input.ConfigPath) {
|
||||
log.Infof("generating new config %s", input.ConfigPath)
|
||||
cfg, err := createNewConfig(input)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
err = util.DirectWriteJson(context.Background(), input.ConfigPath, cfg)
|
||||
return cfg, err
|
||||
}
|
||||
|
||||
if isPreSharedKeyHidden(input.PreSharedKey) {
|
||||
input.PreSharedKey = nil
|
||||
}
|
||||
|
||||
// Enforce permissions on existing config files (same as UpdateOrCreateConfig)
|
||||
if err := util.EnforcePermission(input.ConfigPath); err != nil {
|
||||
log.Errorf("failed to enforce permission on config file: %v", err)
|
||||
}
|
||||
|
||||
return directUpdate(input)
|
||||
}
|
||||
|
||||
func directUpdate(input ConfigInput) (*Config, error) {
|
||||
config := &Config{}
|
||||
|
||||
if _, err := util.ReadJson(input.ConfigPath, config); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
|
||||
updated, err := config.apply(input)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
|
||||
if updated {
|
||||
if err := util.DirectWriteJson(context.Background(), input.ConfigPath, config); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
}
|
||||
|
||||
return config, nil
|
||||
}
|
||||
|
||||
// ConfigToJSON serializes a Config struct to a JSON string.
|
||||
// This is useful for exporting config to alternative storage mechanisms
|
||||
// (e.g., UserDefaults on tvOS where file writes are blocked).
|
||||
func ConfigToJSON(config *Config) (string, error) {
|
||||
bs, err := json.MarshalIndent(config, "", " ")
|
||||
if err != nil {
|
||||
return "", err
|
||||
}
|
||||
return string(bs), nil
|
||||
}
|
||||
|
||||
// ConfigFromJSON deserializes a JSON string to a Config struct.
|
||||
// This is useful for restoring config from alternative storage mechanisms.
|
||||
// After unmarshaling, defaults are applied to ensure the config is fully initialized.
|
||||
func ConfigFromJSON(jsonStr string) (*Config, error) {
|
||||
config := &Config{}
|
||||
err := json.Unmarshal([]byte(jsonStr), config)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
|
||||
// Apply defaults to ensure required fields are initialized.
|
||||
// This mirrors what readConfig does after loading from file.
|
||||
if _, err := config.apply(ConfigInput{}); err != nil {
|
||||
return nil, fmt.Errorf("failed to apply defaults to config: %w", err)
|
||||
}
|
||||
|
||||
return config, nil
|
||||
}
|
||||
|
||||
@@ -76,7 +76,6 @@ func (a *ActiveProfileState) FilePath() (string, error) {
|
||||
}
|
||||
|
||||
type ServiceManager struct {
|
||||
profilesDir string // If set, overrides ConfigDirOverride for profile operations
|
||||
}
|
||||
|
||||
func NewServiceManager(defaultConfigPath string) *ServiceManager {
|
||||
@@ -86,17 +85,6 @@ func NewServiceManager(defaultConfigPath string) *ServiceManager {
|
||||
return &ServiceManager{}
|
||||
}
|
||||
|
||||
// NewServiceManagerWithProfilesDir creates a ServiceManager with a specific profiles directory
|
||||
// This allows setting the profiles directory without modifying the global ConfigDirOverride
|
||||
func NewServiceManagerWithProfilesDir(defaultConfigPath string, profilesDir string) *ServiceManager {
|
||||
if defaultConfigPath != "" {
|
||||
DefaultConfigPath = defaultConfigPath
|
||||
}
|
||||
return &ServiceManager{
|
||||
profilesDir: profilesDir,
|
||||
}
|
||||
}
|
||||
|
||||
func (s *ServiceManager) CopyDefaultProfileIfNotExists() (bool, error) {
|
||||
|
||||
if err := os.MkdirAll(DefaultConfigPathDir, 0600); err != nil {
|
||||
@@ -126,6 +114,14 @@ func (s *ServiceManager) CopyDefaultProfileIfNotExists() (bool, error) {
|
||||
log.Warnf("failed to set permissions for default profile: %v", err)
|
||||
}
|
||||
|
||||
if err := s.SetActiveProfileState(&ActiveProfileState{
|
||||
Name: "default",
|
||||
Username: "",
|
||||
}); err != nil {
|
||||
log.Errorf("failed to set active profile state: %v", err)
|
||||
return false, fmt.Errorf("failed to set active profile state: %w", err)
|
||||
}
|
||||
|
||||
return true, nil
|
||||
}
|
||||
|
||||
@@ -244,7 +240,7 @@ func (s *ServiceManager) DefaultProfilePath() string {
|
||||
}
|
||||
|
||||
func (s *ServiceManager) AddProfile(profileName, username string) error {
|
||||
configDir, err := s.getConfigDir(username)
|
||||
configDir, err := getConfigDirForUser(username)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to get config directory: %w", err)
|
||||
}
|
||||
@@ -274,7 +270,7 @@ func (s *ServiceManager) AddProfile(profileName, username string) error {
|
||||
}
|
||||
|
||||
func (s *ServiceManager) RemoveProfile(profileName, username string) error {
|
||||
configDir, err := s.getConfigDir(username)
|
||||
configDir, err := getConfigDirForUser(username)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to get config directory: %w", err)
|
||||
}
|
||||
@@ -306,7 +302,7 @@ func (s *ServiceManager) RemoveProfile(profileName, username string) error {
|
||||
}
|
||||
|
||||
func (s *ServiceManager) ListProfiles(username string) ([]Profile, error) {
|
||||
configDir, err := s.getConfigDir(username)
|
||||
configDir, err := getConfigDirForUser(username)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("failed to get config directory: %w", err)
|
||||
}
|
||||
@@ -365,7 +361,7 @@ func (s *ServiceManager) GetStatePath() string {
|
||||
return defaultStatePath
|
||||
}
|
||||
|
||||
configDir, err := s.getConfigDir(activeProf.Username)
|
||||
configDir, err := getConfigDirForUser(activeProf.Username)
|
||||
if err != nil {
|
||||
log.Warnf("failed to get config directory for user %s: %v", activeProf.Username, err)
|
||||
return defaultStatePath
|
||||
@@ -373,12 +369,3 @@ func (s *ServiceManager) GetStatePath() string {
|
||||
|
||||
return filepath.Join(configDir, activeProf.Name+".state.json")
|
||||
}
|
||||
|
||||
// getConfigDir returns the profiles directory, using profilesDir if set, otherwise getConfigDirForUser
|
||||
func (s *ServiceManager) getConfigDir(username string) (string, error) {
|
||||
if s.profilesDir != "" {
|
||||
return s.profilesDir, nil
|
||||
}
|
||||
|
||||
return getConfigDirForUser(username)
|
||||
}
|
||||
|
||||
@@ -17,13 +17,12 @@ import (
|
||||
|
||||
nberrors "github.com/netbirdio/netbird/client/errors"
|
||||
firewall "github.com/netbirdio/netbird/client/firewall/manager"
|
||||
"github.com/netbirdio/netbird/client/iface/wgaddr"
|
||||
nbdns "github.com/netbirdio/netbird/client/internal/dns"
|
||||
"github.com/netbirdio/netbird/client/internal/dns/resutil"
|
||||
"github.com/netbirdio/netbird/client/internal/peer"
|
||||
"github.com/netbirdio/netbird/client/internal/peerstore"
|
||||
"github.com/netbirdio/netbird/client/internal/routemanager/common"
|
||||
"github.com/netbirdio/netbird/client/internal/routemanager/fakeip"
|
||||
iface "github.com/netbirdio/netbird/client/internal/routemanager/iface"
|
||||
"github.com/netbirdio/netbird/client/internal/routemanager/refcounter"
|
||||
"github.com/netbirdio/netbird/route"
|
||||
"github.com/netbirdio/netbird/shared/management/domain"
|
||||
@@ -38,6 +37,11 @@ type internalDNATer interface {
|
||||
AddInternalDNATMapping(netip.Addr, netip.Addr) error
|
||||
}
|
||||
|
||||
type wgInterface interface {
|
||||
Name() string
|
||||
Address() wgaddr.Address
|
||||
}
|
||||
|
||||
type DnsInterceptor struct {
|
||||
mu sync.RWMutex
|
||||
route *route.Route
|
||||
@@ -47,7 +51,7 @@ type DnsInterceptor struct {
|
||||
dnsServer nbdns.Server
|
||||
currentPeerKey string
|
||||
interceptedDomains domainMap
|
||||
wgInterface iface.WGIface
|
||||
wgInterface wgInterface
|
||||
peerStore *peerstore.Store
|
||||
firewall firewall.Manager
|
||||
fakeIPManager *fakeip.Manager
|
||||
@@ -215,14 +219,14 @@ func (d *DnsInterceptor) RemoveAllowedIPs() error {
|
||||
|
||||
// ServeDNS implements the dns.Handler interface
|
||||
func (d *DnsInterceptor) ServeDNS(w dns.ResponseWriter, r *dns.Msg) {
|
||||
logger := log.WithFields(log.Fields{
|
||||
"request_id": resutil.GetRequestID(w),
|
||||
"dns_id": fmt.Sprintf("%04x", r.Id),
|
||||
})
|
||||
requestID := nbdns.GenerateRequestID()
|
||||
logger := log.WithField("request_id", requestID)
|
||||
|
||||
if len(r.Question) == 0 {
|
||||
return
|
||||
}
|
||||
logger.Tracef("received DNS request for domain=%s type=%v class=%v",
|
||||
r.Question[0].Name, r.Question[0].Qtype, r.Question[0].Qclass)
|
||||
|
||||
// pass if non A/AAAA query
|
||||
if r.Question[0].Qtype != dns.TypeA && r.Question[0].Qtype != dns.TypeAAAA {
|
||||
@@ -245,6 +249,12 @@ func (d *DnsInterceptor) ServeDNS(w dns.ResponseWriter, r *dns.Msg) {
|
||||
return
|
||||
}
|
||||
|
||||
client, err := nbdns.GetClientPrivate(d.wgInterface.Address().IP, d.wgInterface.Name(), dnsTimeout)
|
||||
if err != nil {
|
||||
d.writeDNSError(w, r, logger, fmt.Sprintf("create DNS client: %v", err))
|
||||
return
|
||||
}
|
||||
|
||||
if r.Extra == nil {
|
||||
r.MsgHdr.AuthenticatedData = true
|
||||
}
|
||||
@@ -253,15 +263,32 @@ func (d *DnsInterceptor) ServeDNS(w dns.ResponseWriter, r *dns.Msg) {
|
||||
ctx, cancel := context.WithTimeout(context.Background(), dnsTimeout)
|
||||
defer cancel()
|
||||
|
||||
reply := d.queryUpstreamDNS(ctx, w, r, upstream, upstreamIP, peerKey, logger)
|
||||
if reply == nil {
|
||||
startTime := time.Now()
|
||||
reply, _, err := nbdns.ExchangeWithFallback(ctx, client, r, upstream)
|
||||
if err != nil {
|
||||
if errors.Is(err, context.DeadlineExceeded) {
|
||||
elapsed := time.Since(startTime)
|
||||
peerInfo := d.debugPeerTimeout(upstreamIP, peerKey)
|
||||
logger.Errorf("peer DNS timeout after %v (timeout=%v) for domain=%s to peer %s (%s)%s - error: %v",
|
||||
elapsed.Truncate(time.Millisecond), dnsTimeout, r.Question[0].Name, upstreamIP.String(), peerKey, peerInfo, err)
|
||||
} else {
|
||||
logger.Errorf("failed to exchange DNS request with %s (%s) for domain=%s: %v", upstreamIP.String(), peerKey, r.Question[0].Name, err)
|
||||
}
|
||||
if err := w.WriteMsg(&dns.Msg{MsgHdr: dns.MsgHdr{Rcode: dns.RcodeServerFailure, Id: r.Id}}); err != nil {
|
||||
logger.Errorf("failed writing DNS response: %v", err)
|
||||
}
|
||||
return
|
||||
}
|
||||
|
||||
resutil.SetMeta(w, "peer", peerKey)
|
||||
var answer []dns.RR
|
||||
if reply != nil {
|
||||
answer = reply.Answer
|
||||
}
|
||||
|
||||
logger.Tracef("upstream %s (%s) DNS response for domain=%s answers=%v", upstreamIP.String(), peerKey, r.Question[0].Name, answer)
|
||||
|
||||
reply.Id = r.Id
|
||||
if err := d.writeMsg(w, reply, logger); err != nil {
|
||||
if err := d.writeMsg(w, reply); err != nil {
|
||||
logger.Errorf("failed writing DNS response: %v", err)
|
||||
}
|
||||
}
|
||||
@@ -297,15 +324,11 @@ func (d *DnsInterceptor) getUpstreamIP(peerKey string) (netip.Addr, error) {
|
||||
return peerAllowedIP, nil
|
||||
}
|
||||
|
||||
func (d *DnsInterceptor) writeMsg(w dns.ResponseWriter, r *dns.Msg, logger *log.Entry) error {
|
||||
func (d *DnsInterceptor) writeMsg(w dns.ResponseWriter, r *dns.Msg) error {
|
||||
if r == nil {
|
||||
return fmt.Errorf("received nil DNS message")
|
||||
}
|
||||
|
||||
// Clear Zero bit from peer responses to prevent external sources from
|
||||
// manipulating our internal fallthrough signaling mechanism
|
||||
r.MsgHdr.Zero = false
|
||||
|
||||
if len(r.Answer) > 0 && len(r.Question) > 0 {
|
||||
origPattern := ""
|
||||
if writer, ok := w.(*nbdns.ResponseWriterChain); ok {
|
||||
@@ -327,14 +350,14 @@ func (d *DnsInterceptor) writeMsg(w dns.ResponseWriter, r *dns.Msg, logger *log.
|
||||
case *dns.A:
|
||||
addr, ok := netip.AddrFromSlice(rr.A)
|
||||
if !ok {
|
||||
logger.Tracef("failed to convert A record for domain=%s ip=%v", resolvedDomain, rr.A)
|
||||
log.Tracef("failed to convert A record for domain=%s ip=%v", resolvedDomain, rr.A)
|
||||
continue
|
||||
}
|
||||
ip = addr
|
||||
case *dns.AAAA:
|
||||
addr, ok := netip.AddrFromSlice(rr.AAAA)
|
||||
if !ok {
|
||||
logger.Tracef("failed to convert AAAA record for domain=%s ip=%v", resolvedDomain, rr.AAAA)
|
||||
log.Tracef("failed to convert AAAA record for domain=%s ip=%v", resolvedDomain, rr.AAAA)
|
||||
continue
|
||||
}
|
||||
ip = addr
|
||||
@@ -347,11 +370,11 @@ func (d *DnsInterceptor) writeMsg(w dns.ResponseWriter, r *dns.Msg, logger *log.
|
||||
}
|
||||
|
||||
if len(newPrefixes) > 0 {
|
||||
if err := d.updateDomainPrefixes(resolvedDomain, originalDomain, newPrefixes, logger); err != nil {
|
||||
logger.Errorf("failed to update domain prefixes: %v", err)
|
||||
if err := d.updateDomainPrefixes(resolvedDomain, originalDomain, newPrefixes); err != nil {
|
||||
log.Errorf("failed to update domain prefixes: %v", err)
|
||||
}
|
||||
|
||||
d.replaceIPsInDNSResponse(r, newPrefixes, logger)
|
||||
d.replaceIPsInDNSResponse(r, newPrefixes)
|
||||
}
|
||||
}
|
||||
|
||||
@@ -363,22 +386,22 @@ func (d *DnsInterceptor) writeMsg(w dns.ResponseWriter, r *dns.Msg, logger *log.
|
||||
}
|
||||
|
||||
// logPrefixChanges handles the logging for prefix changes
|
||||
func (d *DnsInterceptor) logPrefixChanges(resolvedDomain, originalDomain domain.Domain, toAdd, toRemove []netip.Prefix, logger *log.Entry) {
|
||||
func (d *DnsInterceptor) logPrefixChanges(resolvedDomain, originalDomain domain.Domain, toAdd, toRemove []netip.Prefix) {
|
||||
if len(toAdd) > 0 {
|
||||
logger.Debugf("added dynamic route(s) for domain=%s (pattern: domain=%s): %s",
|
||||
log.Debugf("added dynamic route(s) for domain=%s (pattern: domain=%s): %s",
|
||||
resolvedDomain.SafeString(),
|
||||
originalDomain.SafeString(),
|
||||
toAdd)
|
||||
}
|
||||
if len(toRemove) > 0 && !d.route.KeepRoute {
|
||||
logger.Debugf("removed dynamic route(s) for domain=%s (pattern: domain=%s): %s",
|
||||
log.Debugf("removed dynamic route(s) for domain=%s (pattern: domain=%s): %s",
|
||||
resolvedDomain.SafeString(),
|
||||
originalDomain.SafeString(),
|
||||
toRemove)
|
||||
}
|
||||
}
|
||||
|
||||
func (d *DnsInterceptor) updateDomainPrefixes(resolvedDomain, originalDomain domain.Domain, newPrefixes []netip.Prefix, logger *log.Entry) error {
|
||||
func (d *DnsInterceptor) updateDomainPrefixes(resolvedDomain, originalDomain domain.Domain, newPrefixes []netip.Prefix) error {
|
||||
d.mu.Lock()
|
||||
defer d.mu.Unlock()
|
||||
|
||||
@@ -395,9 +418,9 @@ func (d *DnsInterceptor) updateDomainPrefixes(resolvedDomain, originalDomain dom
|
||||
realIP := prefix.Addr()
|
||||
if fakeIP, err := d.fakeIPManager.AllocateFakeIP(realIP); err == nil {
|
||||
dnatMappings[fakeIP] = realIP
|
||||
logger.Tracef("allocated fake IP %s for real IP %s", fakeIP, realIP)
|
||||
log.Tracef("allocated fake IP %s for real IP %s", fakeIP, realIP)
|
||||
} else {
|
||||
logger.Errorf("failed to allocate fake IP for %s: %v", realIP, err)
|
||||
log.Errorf("Failed to allocate fake IP for %s: %v", realIP, err)
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -409,7 +432,7 @@ func (d *DnsInterceptor) updateDomainPrefixes(resolvedDomain, originalDomain dom
|
||||
}
|
||||
}
|
||||
|
||||
d.addDNATMappings(dnatMappings, logger)
|
||||
d.addDNATMappings(dnatMappings)
|
||||
|
||||
if !d.route.KeepRoute {
|
||||
// Remove old prefixes
|
||||
@@ -425,7 +448,7 @@ func (d *DnsInterceptor) updateDomainPrefixes(resolvedDomain, originalDomain dom
|
||||
}
|
||||
}
|
||||
|
||||
d.removeDNATMappings(toRemove, logger)
|
||||
d.removeDNATMappings(toRemove)
|
||||
}
|
||||
|
||||
// Update domain prefixes using resolved domain as key - store real IPs
|
||||
@@ -440,14 +463,14 @@ func (d *DnsInterceptor) updateDomainPrefixes(resolvedDomain, originalDomain dom
|
||||
// Store real IPs for status (user-facing), not fake IPs
|
||||
d.statusRecorder.UpdateResolvedDomainsStates(originalDomain, resolvedDomain, newPrefixes, d.route.GetResourceID())
|
||||
|
||||
d.logPrefixChanges(resolvedDomain, originalDomain, toAdd, toRemove, logger)
|
||||
d.logPrefixChanges(resolvedDomain, originalDomain, toAdd, toRemove)
|
||||
}
|
||||
|
||||
return nberrors.FormatErrorOrNil(merr)
|
||||
}
|
||||
|
||||
// removeDNATMappings removes DNAT mappings from the firewall for real IP prefixes
|
||||
func (d *DnsInterceptor) removeDNATMappings(realPrefixes []netip.Prefix, logger *log.Entry) {
|
||||
func (d *DnsInterceptor) removeDNATMappings(realPrefixes []netip.Prefix) {
|
||||
if len(realPrefixes) == 0 {
|
||||
return
|
||||
}
|
||||
@@ -461,9 +484,9 @@ func (d *DnsInterceptor) removeDNATMappings(realPrefixes []netip.Prefix, logger
|
||||
realIP := prefix.Addr()
|
||||
if fakeIP, exists := d.fakeIPManager.GetFakeIP(realIP); exists {
|
||||
if err := dnatFirewall.RemoveInternalDNATMapping(fakeIP); err != nil {
|
||||
logger.Errorf("failed to remove DNAT mapping for %s: %v", fakeIP, err)
|
||||
log.Errorf("Failed to remove DNAT mapping for %s: %v", fakeIP, err)
|
||||
} else {
|
||||
logger.Debugf("removed DNAT mapping: %s -> %s", fakeIP, realIP)
|
||||
log.Debugf("Removed DNAT mapping for: %s -> %s", fakeIP, realIP)
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -479,7 +502,7 @@ func (d *DnsInterceptor) internalDnatFw() (internalDNATer, bool) {
|
||||
}
|
||||
|
||||
// addDNATMappings adds DNAT mappings to the firewall
|
||||
func (d *DnsInterceptor) addDNATMappings(mappings map[netip.Addr]netip.Addr, logger *log.Entry) {
|
||||
func (d *DnsInterceptor) addDNATMappings(mappings map[netip.Addr]netip.Addr) {
|
||||
if len(mappings) == 0 {
|
||||
return
|
||||
}
|
||||
@@ -491,9 +514,9 @@ func (d *DnsInterceptor) addDNATMappings(mappings map[netip.Addr]netip.Addr, log
|
||||
|
||||
for fakeIP, realIP := range mappings {
|
||||
if err := dnatFirewall.AddInternalDNATMapping(fakeIP, realIP); err != nil {
|
||||
logger.Errorf("failed to add DNAT mapping %s -> %s: %v", fakeIP, realIP, err)
|
||||
log.Errorf("Failed to add DNAT mapping %s -> %s: %v", fakeIP, realIP, err)
|
||||
} else {
|
||||
logger.Debugf("added DNAT mapping: %s -> %s", fakeIP, realIP)
|
||||
log.Debugf("Added DNAT mapping: %s -> %s", fakeIP, realIP)
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -505,12 +528,12 @@ func (d *DnsInterceptor) cleanupDNATMappings() {
|
||||
}
|
||||
|
||||
for _, prefixes := range d.interceptedDomains {
|
||||
d.removeDNATMappings(prefixes, log.NewEntry(log.StandardLogger()))
|
||||
d.removeDNATMappings(prefixes)
|
||||
}
|
||||
}
|
||||
|
||||
// replaceIPsInDNSResponse replaces real IPs with fake IPs in the DNS response
|
||||
func (d *DnsInterceptor) replaceIPsInDNSResponse(reply *dns.Msg, realPrefixes []netip.Prefix, logger *log.Entry) {
|
||||
func (d *DnsInterceptor) replaceIPsInDNSResponse(reply *dns.Msg, realPrefixes []netip.Prefix) {
|
||||
if _, ok := d.internalDnatFw(); !ok {
|
||||
return
|
||||
}
|
||||
@@ -526,7 +549,7 @@ func (d *DnsInterceptor) replaceIPsInDNSResponse(reply *dns.Msg, realPrefixes []
|
||||
|
||||
if fakeIP, exists := d.fakeIPManager.GetFakeIP(realIP); exists {
|
||||
rr.A = fakeIP.AsSlice()
|
||||
logger.Tracef("replaced real IP %s with fake IP %s in DNS response", realIP, fakeIP)
|
||||
log.Tracef("Replaced real IP %s with fake IP %s in DNS response", realIP, fakeIP)
|
||||
}
|
||||
|
||||
case *dns.AAAA:
|
||||
@@ -537,7 +560,7 @@ func (d *DnsInterceptor) replaceIPsInDNSResponse(reply *dns.Msg, realPrefixes []
|
||||
|
||||
if fakeIP, exists := d.fakeIPManager.GetFakeIP(realIP); exists {
|
||||
rr.AAAA = fakeIP.AsSlice()
|
||||
logger.Tracef("replaced real IP %s with fake IP %s in DNS response", realIP, fakeIP)
|
||||
log.Tracef("Replaced real IP %s with fake IP %s in DNS response", realIP, fakeIP)
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -563,44 +586,6 @@ func determinePrefixChanges(oldPrefixes, newPrefixes []netip.Prefix) (toAdd, toR
|
||||
return
|
||||
}
|
||||
|
||||
// queryUpstreamDNS queries the upstream DNS server using netstack if available, otherwise uses regular client.
|
||||
// Returns the DNS reply on success, or nil on error (error responses are written internally).
|
||||
func (d *DnsInterceptor) queryUpstreamDNS(ctx context.Context, w dns.ResponseWriter, r *dns.Msg, upstream string, upstreamIP netip.Addr, peerKey string, logger *log.Entry) *dns.Msg {
|
||||
startTime := time.Now()
|
||||
|
||||
nsNet := d.wgInterface.GetNet()
|
||||
var reply *dns.Msg
|
||||
var err error
|
||||
|
||||
if nsNet != nil {
|
||||
reply, err = nbdns.ExchangeWithNetstack(ctx, nsNet, r, upstream)
|
||||
} else {
|
||||
client, clientErr := nbdns.GetClientPrivate(d.wgInterface.Address().IP, d.wgInterface.Name(), dnsTimeout)
|
||||
if clientErr != nil {
|
||||
d.writeDNSError(w, r, logger, fmt.Sprintf("create DNS client: %v", clientErr))
|
||||
return nil
|
||||
}
|
||||
reply, _, err = nbdns.ExchangeWithFallback(ctx, client, r, upstream)
|
||||
}
|
||||
|
||||
if err == nil {
|
||||
return reply
|
||||
}
|
||||
|
||||
if errors.Is(err, context.DeadlineExceeded) {
|
||||
elapsed := time.Since(startTime)
|
||||
peerInfo := d.debugPeerTimeout(upstreamIP, peerKey)
|
||||
logger.Errorf("peer DNS timeout after %v (timeout=%v) for domain=%s to peer %s (%s)%s - error: %v",
|
||||
elapsed.Truncate(time.Millisecond), dnsTimeout, r.Question[0].Name, upstreamIP.String(), peerKey, peerInfo, err)
|
||||
} else {
|
||||
logger.Errorf("failed to exchange DNS request with %s (%s) for domain=%s: %v", upstreamIP.String(), peerKey, r.Question[0].Name, err)
|
||||
}
|
||||
if err := w.WriteMsg(&dns.Msg{MsgHdr: dns.MsgHdr{Rcode: dns.RcodeServerFailure, Id: r.Id}}); err != nil {
|
||||
logger.Errorf("failed writing DNS response: %v", err)
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
func (d *DnsInterceptor) debugPeerTimeout(peerIP netip.Addr, peerKey string) string {
|
||||
if d.statusRecorder == nil {
|
||||
return ""
|
||||
|
||||
@@ -1,4 +1,5 @@
|
||||
//go:build !windows
|
||||
// +build !windows
|
||||
|
||||
package iface
|
||||
|
||||
|
||||
@@ -4,8 +4,6 @@ import (
|
||||
"net"
|
||||
"net/netip"
|
||||
|
||||
"golang.zx2c4.com/wireguard/tun/netstack"
|
||||
|
||||
"github.com/netbirdio/netbird/client/iface/device"
|
||||
"github.com/netbirdio/netbird/client/iface/wgaddr"
|
||||
)
|
||||
@@ -20,5 +18,4 @@ type wgIfaceBase interface {
|
||||
IsUserspaceBind() bool
|
||||
GetFilter() device.PacketFilter
|
||||
GetDevice() *device.FilteredDevice
|
||||
GetNet() *netstack.Net
|
||||
}
|
||||
|
||||
@@ -210,8 +210,7 @@ func (r *SysOps) refreshLocalSubnetsCache() {
|
||||
func (r *SysOps) genericAddVPNRoute(prefix netip.Prefix, intf *net.Interface) error {
|
||||
nextHop := Nexthop{netip.Addr{}, intf}
|
||||
|
||||
switch prefix {
|
||||
case vars.Defaultv4:
|
||||
if prefix == vars.Defaultv4 {
|
||||
if err := r.addToRouteTable(splitDefaultv4_1, nextHop); err != nil {
|
||||
return err
|
||||
}
|
||||
@@ -234,7 +233,7 @@ func (r *SysOps) genericAddVPNRoute(prefix netip.Prefix, intf *net.Interface) er
|
||||
}
|
||||
|
||||
return nil
|
||||
case vars.Defaultv6:
|
||||
} else if prefix == vars.Defaultv6 {
|
||||
if err := r.addToRouteTable(splitDefaultv6_1, nextHop); err != nil {
|
||||
return fmt.Errorf("add unreachable route split 1: %w", err)
|
||||
}
|
||||
@@ -256,8 +255,7 @@ func (r *SysOps) genericAddVPNRoute(prefix netip.Prefix, intf *net.Interface) er
|
||||
func (r *SysOps) genericRemoveVPNRoute(prefix netip.Prefix, intf *net.Interface) error {
|
||||
nextHop := Nexthop{netip.Addr{}, intf}
|
||||
|
||||
switch prefix {
|
||||
case vars.Defaultv4:
|
||||
if prefix == vars.Defaultv4 {
|
||||
var result *multierror.Error
|
||||
if err := r.removeFromRouteTable(splitDefaultv4_1, nextHop); err != nil {
|
||||
result = multierror.Append(result, err)
|
||||
@@ -275,7 +273,7 @@ func (r *SysOps) genericRemoveVPNRoute(prefix netip.Prefix, intf *net.Interface)
|
||||
}
|
||||
|
||||
return nberrors.FormatErrorOrNil(result)
|
||||
case vars.Defaultv6:
|
||||
} else if prefix == vars.Defaultv6 {
|
||||
var result *multierror.Error
|
||||
if err := r.removeFromRouteTable(splitDefaultv6_1, nextHop); err != nil {
|
||||
result = multierror.Append(result, err)
|
||||
@@ -285,9 +283,9 @@ func (r *SysOps) genericRemoveVPNRoute(prefix netip.Prefix, intf *net.Interface)
|
||||
}
|
||||
|
||||
return nberrors.FormatErrorOrNil(result)
|
||||
default:
|
||||
return r.removeFromRouteTable(prefix, nextHop)
|
||||
}
|
||||
|
||||
return r.removeFromRouteTable(prefix, nextHop)
|
||||
}
|
||||
|
||||
func (r *SysOps) setupHooks(initAddresses []net.IP, stateManager *statemanager.Manager) error {
|
||||
|
||||
@@ -1,35 +0,0 @@
|
||||
// Package updatemanager provides automatic update management for the NetBird client.
|
||||
// It monitors for new versions, handles update triggers from management server directives,
|
||||
// and orchestrates the download and installation of client updates.
|
||||
//
|
||||
// # Overview
|
||||
//
|
||||
// The update manager operates as a background service that continuously monitors for
|
||||
// available updates and automatically initiates the update process when conditions are met.
|
||||
// It integrates with the installer package to perform the actual installation.
|
||||
//
|
||||
// # Update Flow
|
||||
//
|
||||
// The complete update process follows these steps:
|
||||
//
|
||||
// 1. Manager receives update directive via SetVersion() or detects new version
|
||||
// 2. Manager validates update should proceed (version comparison, rate limiting)
|
||||
// 3. Manager publishes "updating" event to status recorder
|
||||
// 4. Manager persists UpdateState to track update attempt
|
||||
// 5. Manager downloads installer file (.msi or .exe) to temporary directory
|
||||
// 6. Manager triggers installation via installer.RunInstallation()
|
||||
// 7. Installer package handles the actual installation process
|
||||
// 8. On next startup, CheckUpdateSuccess() verifies update completion
|
||||
// 9. Manager publishes success/failure event to status recorder
|
||||
// 10. Manager cleans up UpdateState
|
||||
//
|
||||
// # State Management
|
||||
//
|
||||
// Update state is persisted across restarts to track update attempts:
|
||||
//
|
||||
// - PreUpdateVersion: Version before update attempt
|
||||
// - TargetVersion: Version attempting to update to
|
||||
//
|
||||
// This enables verification of successful updates and appropriate user notification
|
||||
// after the client restarts with the new version.
|
||||
package updatemanager
|
||||
@@ -1,138 +0,0 @@
|
||||
package downloader
|
||||
|
||||
import (
|
||||
"context"
|
||||
"fmt"
|
||||
"io"
|
||||
"net/http"
|
||||
"os"
|
||||
"time"
|
||||
|
||||
log "github.com/sirupsen/logrus"
|
||||
|
||||
"github.com/netbirdio/netbird/version"
|
||||
)
|
||||
|
||||
const (
|
||||
userAgent = "NetBird agent installer/%s"
|
||||
DefaultRetryDelay = 3 * time.Second
|
||||
)
|
||||
|
||||
func DownloadToFile(ctx context.Context, retryDelay time.Duration, url, dstFile string) error {
|
||||
log.Debugf("starting download from %s", url)
|
||||
|
||||
out, err := os.Create(dstFile)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to create destination file %q: %w", dstFile, err)
|
||||
}
|
||||
defer func() {
|
||||
if cerr := out.Close(); cerr != nil {
|
||||
log.Warnf("error closing file %q: %v", dstFile, cerr)
|
||||
}
|
||||
}()
|
||||
|
||||
// First attempt
|
||||
err = downloadToFileOnce(ctx, url, out)
|
||||
if err == nil {
|
||||
log.Infof("successfully downloaded file to %s", dstFile)
|
||||
return nil
|
||||
}
|
||||
|
||||
// If retryDelay is 0, don't retry
|
||||
if retryDelay == 0 {
|
||||
return err
|
||||
}
|
||||
|
||||
log.Warnf("download failed, retrying after %v: %v", retryDelay, err)
|
||||
|
||||
// Sleep before retry
|
||||
if sleepErr := sleepWithContext(ctx, retryDelay); sleepErr != nil {
|
||||
return fmt.Errorf("download cancelled during retry delay: %w", sleepErr)
|
||||
}
|
||||
|
||||
// Truncate file before retry
|
||||
if err := out.Truncate(0); err != nil {
|
||||
return fmt.Errorf("failed to truncate file on retry: %w", err)
|
||||
}
|
||||
if _, err := out.Seek(0, 0); err != nil {
|
||||
return fmt.Errorf("failed to seek to beginning of file: %w", err)
|
||||
}
|
||||
|
||||
// Second attempt
|
||||
if err := downloadToFileOnce(ctx, url, out); err != nil {
|
||||
return fmt.Errorf("download failed after retry: %w", err)
|
||||
}
|
||||
|
||||
log.Infof("successfully downloaded file to %s", dstFile)
|
||||
return nil
|
||||
}
|
||||
|
||||
func DownloadToMemory(ctx context.Context, url string, limit int64) ([]byte, error) {
|
||||
req, err := http.NewRequestWithContext(ctx, http.MethodGet, url, nil)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("failed to create HTTP request: %w", err)
|
||||
}
|
||||
|
||||
// Add User-Agent header
|
||||
req.Header.Set("User-Agent", fmt.Sprintf(userAgent, version.NetbirdVersion()))
|
||||
|
||||
resp, err := http.DefaultClient.Do(req)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("failed to perform HTTP request: %w", err)
|
||||
}
|
||||
defer func() {
|
||||
if cerr := resp.Body.Close(); cerr != nil {
|
||||
log.Warnf("error closing response body: %v", cerr)
|
||||
}
|
||||
}()
|
||||
|
||||
if resp.StatusCode != http.StatusOK {
|
||||
return nil, fmt.Errorf("unexpected HTTP status: %d", resp.StatusCode)
|
||||
}
|
||||
|
||||
data, err := io.ReadAll(io.LimitReader(resp.Body, limit))
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("failed to read response body: %w", err)
|
||||
}
|
||||
|
||||
return data, nil
|
||||
}
|
||||
|
||||
func downloadToFileOnce(ctx context.Context, url string, out *os.File) error {
|
||||
req, err := http.NewRequestWithContext(ctx, http.MethodGet, url, nil)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to create HTTP request: %w", err)
|
||||
}
|
||||
|
||||
// Add User-Agent header
|
||||
req.Header.Set("User-Agent", fmt.Sprintf(userAgent, version.NetbirdVersion()))
|
||||
|
||||
resp, err := http.DefaultClient.Do(req)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to perform HTTP request: %w", err)
|
||||
}
|
||||
defer func() {
|
||||
if cerr := resp.Body.Close(); cerr != nil {
|
||||
log.Warnf("error closing response body: %v", cerr)
|
||||
}
|
||||
}()
|
||||
|
||||
if resp.StatusCode != http.StatusOK {
|
||||
return fmt.Errorf("unexpected HTTP status: %d", resp.StatusCode)
|
||||
}
|
||||
|
||||
if _, err := io.Copy(out, resp.Body); err != nil {
|
||||
return fmt.Errorf("failed to write response body to file: %w", err)
|
||||
}
|
||||
|
||||
return nil
|
||||
}
|
||||
|
||||
func sleepWithContext(ctx context.Context, duration time.Duration) error {
|
||||
select {
|
||||
case <-time.After(duration):
|
||||
return nil
|
||||
case <-ctx.Done():
|
||||
return ctx.Err()
|
||||
}
|
||||
}
|
||||
@@ -1,199 +0,0 @@
|
||||
package downloader
|
||||
|
||||
import (
|
||||
"context"
|
||||
"net/http"
|
||||
"net/http/httptest"
|
||||
"os"
|
||||
"path/filepath"
|
||||
"sync/atomic"
|
||||
"testing"
|
||||
"time"
|
||||
)
|
||||
|
||||
const (
|
||||
retryDelay = 100 * time.Millisecond
|
||||
)
|
||||
|
||||
func TestDownloadToFile_Success(t *testing.T) {
|
||||
// Create a test server that responds successfully
|
||||
content := "test file content"
|
||||
server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
w.WriteHeader(http.StatusOK)
|
||||
_, _ = w.Write([]byte(content))
|
||||
}))
|
||||
defer server.Close()
|
||||
|
||||
// Create a temporary file for download
|
||||
tempDir := t.TempDir()
|
||||
dstFile := filepath.Join(tempDir, "downloaded.txt")
|
||||
|
||||
// Download the file
|
||||
err := DownloadToFile(context.Background(), retryDelay, server.URL, dstFile)
|
||||
if err != nil {
|
||||
t.Fatalf("expected no error, got: %v", err)
|
||||
}
|
||||
|
||||
// Verify the file content
|
||||
data, err := os.ReadFile(dstFile)
|
||||
if err != nil {
|
||||
t.Fatalf("failed to read downloaded file: %v", err)
|
||||
}
|
||||
|
||||
if string(data) != content {
|
||||
t.Errorf("expected content %q, got %q", content, string(data))
|
||||
}
|
||||
}
|
||||
|
||||
func TestDownloadToFile_SuccessAfterRetry(t *testing.T) {
|
||||
content := "test file content after retry"
|
||||
var attemptCount atomic.Int32
|
||||
|
||||
// Create a test server that fails on first attempt, succeeds on second
|
||||
server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
attempt := attemptCount.Add(1)
|
||||
if attempt == 1 {
|
||||
w.WriteHeader(http.StatusInternalServerError)
|
||||
_, _ = w.Write([]byte("error"))
|
||||
return
|
||||
}
|
||||
w.WriteHeader(http.StatusOK)
|
||||
_, _ = w.Write([]byte(content))
|
||||
}))
|
||||
defer server.Close()
|
||||
|
||||
// Create a temporary file for download
|
||||
tempDir := t.TempDir()
|
||||
dstFile := filepath.Join(tempDir, "downloaded.txt")
|
||||
|
||||
// Download the file (should succeed after retry)
|
||||
if err := DownloadToFile(context.Background(), 10*time.Millisecond, server.URL, dstFile); err != nil {
|
||||
t.Fatalf("expected no error after retry, got: %v", err)
|
||||
}
|
||||
|
||||
// Verify the file content
|
||||
data, err := os.ReadFile(dstFile)
|
||||
if err != nil {
|
||||
t.Fatalf("failed to read downloaded file: %v", err)
|
||||
}
|
||||
|
||||
if string(data) != content {
|
||||
t.Errorf("expected content %q, got %q", content, string(data))
|
||||
}
|
||||
|
||||
// Verify it took 2 attempts
|
||||
if attemptCount.Load() != 2 {
|
||||
t.Errorf("expected 2 attempts, got %d", attemptCount.Load())
|
||||
}
|
||||
}
|
||||
|
||||
func TestDownloadToFile_FailsAfterRetry(t *testing.T) {
|
||||
var attemptCount atomic.Int32
|
||||
|
||||
// Create a test server that always fails
|
||||
server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
attemptCount.Add(1)
|
||||
w.WriteHeader(http.StatusInternalServerError)
|
||||
_, _ = w.Write([]byte("error"))
|
||||
}))
|
||||
defer server.Close()
|
||||
|
||||
// Create a temporary file for download
|
||||
tempDir := t.TempDir()
|
||||
dstFile := filepath.Join(tempDir, "downloaded.txt")
|
||||
|
||||
// Download the file (should fail after retry)
|
||||
if err := DownloadToFile(context.Background(), 10*time.Millisecond, server.URL, dstFile); err == nil {
|
||||
t.Fatal("expected error after retry, got nil")
|
||||
}
|
||||
|
||||
// Verify it tried 2 times
|
||||
if attemptCount.Load() != 2 {
|
||||
t.Errorf("expected 2 attempts, got %d", attemptCount.Load())
|
||||
}
|
||||
}
|
||||
|
||||
func TestDownloadToFile_ContextCancellationDuringRetry(t *testing.T) {
|
||||
var attemptCount atomic.Int32
|
||||
|
||||
// Create a test server that always fails
|
||||
server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
attemptCount.Add(1)
|
||||
w.WriteHeader(http.StatusInternalServerError)
|
||||
}))
|
||||
defer server.Close()
|
||||
|
||||
// Create a temporary file for download
|
||||
tempDir := t.TempDir()
|
||||
dstFile := filepath.Join(tempDir, "downloaded.txt")
|
||||
|
||||
// Create a context that will be cancelled during retry delay
|
||||
ctx, cancel := context.WithCancel(context.Background())
|
||||
|
||||
// Cancel after a short delay (during the retry sleep)
|
||||
go func() {
|
||||
time.Sleep(100 * time.Millisecond)
|
||||
cancel()
|
||||
}()
|
||||
|
||||
// Download the file (should fail due to context cancellation during retry)
|
||||
err := DownloadToFile(ctx, 1*time.Second, server.URL, dstFile)
|
||||
if err == nil {
|
||||
t.Fatal("expected error due to context cancellation, got nil")
|
||||
}
|
||||
|
||||
// Should have only made 1 attempt (cancelled during retry delay)
|
||||
if attemptCount.Load() != 1 {
|
||||
t.Errorf("expected 1 attempt, got %d", attemptCount.Load())
|
||||
}
|
||||
}
|
||||
|
||||
func TestDownloadToFile_InvalidURL(t *testing.T) {
|
||||
tempDir := t.TempDir()
|
||||
dstFile := filepath.Join(tempDir, "downloaded.txt")
|
||||
|
||||
err := DownloadToFile(context.Background(), retryDelay, "://invalid-url", dstFile)
|
||||
if err == nil {
|
||||
t.Fatal("expected error for invalid URL, got nil")
|
||||
}
|
||||
}
|
||||
|
||||
func TestDownloadToFile_InvalidDestination(t *testing.T) {
|
||||
server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
w.WriteHeader(http.StatusOK)
|
||||
_, _ = w.Write([]byte("test"))
|
||||
}))
|
||||
defer server.Close()
|
||||
|
||||
// Use an invalid destination path
|
||||
err := DownloadToFile(context.Background(), retryDelay, server.URL, "/invalid/path/that/does/not/exist/file.txt")
|
||||
if err == nil {
|
||||
t.Fatal("expected error for invalid destination, got nil")
|
||||
}
|
||||
}
|
||||
|
||||
func TestDownloadToFile_NoRetry(t *testing.T) {
|
||||
var attemptCount atomic.Int32
|
||||
|
||||
// Create a test server that always fails
|
||||
server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
|
||||
attemptCount.Add(1)
|
||||
w.WriteHeader(http.StatusInternalServerError)
|
||||
_, _ = w.Write([]byte("error"))
|
||||
}))
|
||||
defer server.Close()
|
||||
|
||||
// Create a temporary file for download
|
||||
tempDir := t.TempDir()
|
||||
dstFile := filepath.Join(tempDir, "downloaded.txt")
|
||||
|
||||
// Download the file with retryDelay = 0 (should not retry)
|
||||
if err := DownloadToFile(context.Background(), 0, server.URL, dstFile); err == nil {
|
||||
t.Fatal("expected error, got nil")
|
||||
}
|
||||
|
||||
// Verify it only made 1 attempt (no retry)
|
||||
if attemptCount.Load() != 1 {
|
||||
t.Errorf("expected 1 attempt, got %d", attemptCount.Load())
|
||||
}
|
||||
}
|
||||
@@ -1,7 +0,0 @@
|
||||
//go:build !windows
|
||||
|
||||
package installer
|
||||
|
||||
func UpdaterBinaryNameWithoutExtension() string {
|
||||
return updaterBinary
|
||||
}
|
||||
@@ -1,11 +0,0 @@
|
||||
package installer
|
||||
|
||||
import (
|
||||
"path/filepath"
|
||||
"strings"
|
||||
)
|
||||
|
||||
func UpdaterBinaryNameWithoutExtension() string {
|
||||
ext := filepath.Ext(updaterBinary)
|
||||
return strings.TrimSuffix(updaterBinary, ext)
|
||||
}
|
||||
@@ -1,111 +0,0 @@
|
||||
// Package installer provides functionality for managing NetBird application
|
||||
// updates and installations across Windows, macOS. It handles
|
||||
// the complete update lifecycle including artifact download, cryptographic verification,
|
||||
// installation execution, process management, and result reporting.
|
||||
//
|
||||
// # Architecture
|
||||
//
|
||||
// The installer package uses a two-process architecture to enable self-updates:
|
||||
//
|
||||
// 1. Service Process: The main NetBird daemon process that initiates updates
|
||||
// 2. Updater Process: A detached child process that performs the actual installation
|
||||
//
|
||||
// This separation is critical because:
|
||||
// - The service binary cannot update itself while running
|
||||
// - The installer (EXE/MSI/PKG) will terminate the service during installation
|
||||
// - The updater process survives service termination and restarts it after installation
|
||||
// - Results can be communicated back to the service after it restarts
|
||||
//
|
||||
// # Update Flow
|
||||
//
|
||||
// Service Process (RunInstallation):
|
||||
//
|
||||
// 1. Validates target version format (semver)
|
||||
// 2. Determines installer type (EXE, MSI, PKG, or Homebrew)
|
||||
// 3. Downloads installer file from GitHub releases (if applicable)
|
||||
// 4. Verifies installer signature using reposign package (cryptographic verification in service process before
|
||||
// launching updater)
|
||||
// 5. Copies service binary to tempDir as "updater" (or "updater.exe" on Windows)
|
||||
// 6. Launches updater process with detached mode:
|
||||
// - --temp-dir: Temporary directory path
|
||||
// - --service-dir: Service installation directory
|
||||
// - --installer-file: Path to downloaded installer (if applicable)
|
||||
// - --dry-run: Optional flag to test without actually installing
|
||||
// 7. Service process continues running (will be terminated by installer later)
|
||||
// 8. Service can watch for result.json using ResultHandler.Watch() to detect completion
|
||||
//
|
||||
// Updater Process (Setup):
|
||||
//
|
||||
// 1. Receives parameters from service via command-line arguments
|
||||
// 2. Runs installer with appropriate silent/quiet flags:
|
||||
// - Windows EXE: installer.exe /S
|
||||
// - Windows MSI: msiexec.exe /i installer.msi /quiet /qn /l*v msi.log
|
||||
// - macOS PKG: installer -pkg installer.pkg -target /
|
||||
// - macOS Homebrew: brew upgrade netbirdio/tap/netbird
|
||||
// 3. Installer terminates daemon and UI processes
|
||||
// 4. Installer replaces binaries with new version
|
||||
// 5. Updater waits for installer to complete
|
||||
// 6. Updater restarts daemon:
|
||||
// - Windows: netbird.exe service start
|
||||
// - macOS/Linux: netbird service start
|
||||
// 7. Updater restarts UI:
|
||||
// - Windows: Launches netbird-ui.exe as active console user using CreateProcessAsUser
|
||||
// - macOS: Uses launchctl asuser to launch NetBird.app for console user
|
||||
// - Linux: Not implemented (UI typically auto-starts)
|
||||
// 8. Updater writes result.json with success/error status
|
||||
// 9. Updater process exits
|
||||
//
|
||||
// # Result Communication
|
||||
//
|
||||
// The ResultHandler (result.go) manages communication between updater and service:
|
||||
//
|
||||
// Result Structure:
|
||||
//
|
||||
// type Result struct {
|
||||
// Success bool // true if installation succeeded
|
||||
// Error string // error message if Success is false
|
||||
// ExecutedAt time.Time // when installation completed
|
||||
// }
|
||||
//
|
||||
// Result files are automatically cleaned up after being read.
|
||||
//
|
||||
// # File Locations
|
||||
//
|
||||
// Temporary Directory (platform-specific):
|
||||
//
|
||||
// Windows:
|
||||
// - Path: %ProgramData%\Netbird\tmp-install
|
||||
// - Example: C:\ProgramData\Netbird\tmp-install
|
||||
//
|
||||
// macOS:
|
||||
// - Path: /var/lib/netbird/tmp-install
|
||||
// - Requires root permissions
|
||||
//
|
||||
// Files created during installation:
|
||||
//
|
||||
// tmp-install/
|
||||
// installer.log
|
||||
// updater[.exe] # Copy of service binary
|
||||
// netbird_installer_*.[exe|msi|pkg] # Downloaded installer
|
||||
// result.json # Installation result
|
||||
// msi.log # MSI verbose log (Windows MSI only)
|
||||
//
|
||||
// # API Reference
|
||||
//
|
||||
// # Cleanup
|
||||
//
|
||||
// CleanUpInstallerFiles() removes temporary files after successful installation:
|
||||
// - Downloaded installer files (*.exe, *.msi, *.pkg)
|
||||
// - Updater binary copy
|
||||
// - Does NOT remove result.json (cleaned by ResultHandler after read)
|
||||
// - Does NOT remove msi.log (kept for debugging)
|
||||
//
|
||||
// # Dry-Run Mode
|
||||
//
|
||||
// Dry-run mode allows testing the update process without actually installing:
|
||||
//
|
||||
// Enable via environment variable:
|
||||
//
|
||||
// export NB_AUTO_UPDATE_DRY_RUN=true
|
||||
// netbird service install-update 0.29.0
|
||||
package installer
|
||||
@@ -1,50 +0,0 @@
|
||||
//go:build !windows && !darwin
|
||||
|
||||
package installer
|
||||
|
||||
import (
|
||||
"context"
|
||||
"fmt"
|
||||
)
|
||||
|
||||
const (
|
||||
updaterBinary = "updater"
|
||||
)
|
||||
|
||||
type Installer struct {
|
||||
tempDir string
|
||||
}
|
||||
|
||||
// New used by the service
|
||||
func New() *Installer {
|
||||
return &Installer{}
|
||||
}
|
||||
|
||||
// NewWithDir used by the updater process, get the tempDir from the service via cmd line
|
||||
func NewWithDir(tempDir string) *Installer {
|
||||
return &Installer{
|
||||
tempDir: tempDir,
|
||||
}
|
||||
}
|
||||
|
||||
func (u *Installer) TempDir() string {
|
||||
return ""
|
||||
}
|
||||
|
||||
func (c *Installer) LogFiles() []string {
|
||||
return []string{}
|
||||
}
|
||||
|
||||
func (u *Installer) CleanUpInstallerFiles() error {
|
||||
return nil
|
||||
}
|
||||
|
||||
func (u *Installer) RunInstallation(ctx context.Context, targetVersion string) error {
|
||||
return fmt.Errorf("unsupported platform")
|
||||
}
|
||||
|
||||
// Setup runs the installer with appropriate arguments and manages the daemon/UI state
|
||||
// This will be run by the updater process
|
||||
func (u *Installer) Setup(ctx context.Context, dryRun bool, targetVersion string, daemonFolder string) (resultErr error) {
|
||||
return fmt.Errorf("unsupported platform")
|
||||
}
|
||||
@@ -1,293 +0,0 @@
|
||||
//go:build windows || darwin
|
||||
|
||||
package installer
|
||||
|
||||
import (
|
||||
"context"
|
||||
"fmt"
|
||||
"io"
|
||||
"os"
|
||||
"os/exec"
|
||||
"path"
|
||||
"path/filepath"
|
||||
"strings"
|
||||
|
||||
"github.com/hashicorp/go-multierror"
|
||||
goversion "github.com/hashicorp/go-version"
|
||||
log "github.com/sirupsen/logrus"
|
||||
|
||||
"github.com/netbirdio/netbird/client/internal/updatemanager/downloader"
|
||||
"github.com/netbirdio/netbird/client/internal/updatemanager/reposign"
|
||||
)
|
||||
|
||||
type Installer struct {
|
||||
tempDir string
|
||||
}
|
||||
|
||||
// New used by the service
|
||||
func New() *Installer {
|
||||
return &Installer{
|
||||
tempDir: defaultTempDir,
|
||||
}
|
||||
}
|
||||
|
||||
// NewWithDir used by the updater process, get the tempDir from the service via cmd line
|
||||
func NewWithDir(tempDir string) *Installer {
|
||||
return &Installer{
|
||||
tempDir: tempDir,
|
||||
}
|
||||
}
|
||||
|
||||
// RunInstallation starts the updater process to run the installation
|
||||
// This will run by the original service process
|
||||
func (u *Installer) RunInstallation(ctx context.Context, targetVersion string) (err error) {
|
||||
resultHandler := NewResultHandler(u.tempDir)
|
||||
|
||||
defer func() {
|
||||
if err != nil {
|
||||
if writeErr := resultHandler.WriteErr(err); writeErr != nil {
|
||||
log.Errorf("failed to write error result: %v", writeErr)
|
||||
}
|
||||
}
|
||||
}()
|
||||
|
||||
if err := validateTargetVersion(targetVersion); err != nil {
|
||||
return err
|
||||
}
|
||||
|
||||
if err := u.mkTempDir(); err != nil {
|
||||
return err
|
||||
}
|
||||
|
||||
var installerFile string
|
||||
// Download files only when not using any third-party store
|
||||
if installerType := TypeOfInstaller(ctx); installerType.Downloadable() {
|
||||
log.Infof("download installer")
|
||||
var err error
|
||||
installerFile, err = u.downloadInstaller(ctx, installerType, targetVersion)
|
||||
if err != nil {
|
||||
log.Errorf("failed to download installer: %v", err)
|
||||
return err
|
||||
}
|
||||
|
||||
artifactVerify, err := reposign.NewArtifactVerify(DefaultSigningKeysBaseURL)
|
||||
if err != nil {
|
||||
log.Errorf("failed to create artifact verify: %v", err)
|
||||
return err
|
||||
}
|
||||
|
||||
if err := artifactVerify.Verify(ctx, targetVersion, installerFile); err != nil {
|
||||
log.Errorf("artifact verification error: %v", err)
|
||||
return err
|
||||
}
|
||||
}
|
||||
|
||||
log.Infof("running installer")
|
||||
updaterPath, err := u.copyUpdater()
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
|
||||
// the directory where the service has been installed
|
||||
workspace, err := getServiceDir()
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
|
||||
args := []string{
|
||||
"--temp-dir", u.tempDir,
|
||||
"--service-dir", workspace,
|
||||
}
|
||||
|
||||
if isDryRunEnabled() {
|
||||
args = append(args, "--dry-run=true")
|
||||
}
|
||||
|
||||
if installerFile != "" {
|
||||
args = append(args, "--installer-file", installerFile)
|
||||
}
|
||||
|
||||
updateCmd := exec.Command(updaterPath, args...)
|
||||
log.Infof("starting updater process: %s", updateCmd.String())
|
||||
|
||||
// Configure the updater to run in a separate session/process group
|
||||
// so it survives the parent daemon being stopped
|
||||
setUpdaterProcAttr(updateCmd)
|
||||
|
||||
// Start the updater process asynchronously
|
||||
if err := updateCmd.Start(); err != nil {
|
||||
return err
|
||||
}
|
||||
|
||||
pid := updateCmd.Process.Pid
|
||||
log.Infof("updater started with PID %d", pid)
|
||||
|
||||
// Release the process so the OS can fully detach it
|
||||
if err := updateCmd.Process.Release(); err != nil {
|
||||
log.Warnf("failed to release updater process: %v", err)
|
||||
}
|
||||
|
||||
return nil
|
||||
}
|
||||
|
||||
// CleanUpInstallerFiles
|
||||
// - the installer file (pkg, exe, msi)
|
||||
// - the selfcopy updater.exe
|
||||
func (u *Installer) CleanUpInstallerFiles() error {
|
||||
// Check if tempDir exists
|
||||
info, err := os.Stat(u.tempDir)
|
||||
if err != nil {
|
||||
if os.IsNotExist(err) {
|
||||
return nil
|
||||
}
|
||||
return err
|
||||
}
|
||||
|
||||
if !info.IsDir() {
|
||||
return nil
|
||||
}
|
||||
|
||||
var merr *multierror.Error
|
||||
|
||||
if err := os.Remove(filepath.Join(u.tempDir, updaterBinary)); err != nil && !os.IsNotExist(err) {
|
||||
merr = multierror.Append(merr, fmt.Errorf("failed to remove updater binary: %w", err))
|
||||
}
|
||||
|
||||
entries, err := os.ReadDir(u.tempDir)
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
|
||||
for _, entry := range entries {
|
||||
if entry.IsDir() {
|
||||
continue
|
||||
}
|
||||
|
||||
name := entry.Name()
|
||||
for _, ext := range binaryExtensions {
|
||||
if strings.HasSuffix(strings.ToLower(name), strings.ToLower(ext)) {
|
||||
if err := os.Remove(filepath.Join(u.tempDir, name)); err != nil {
|
||||
merr = multierror.Append(merr, fmt.Errorf("failed to remove %s: %w", name, err))
|
||||
}
|
||||
break
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
return merr.ErrorOrNil()
|
||||
}
|
||||
|
||||
func (u *Installer) downloadInstaller(ctx context.Context, installerType Type, targetVersion string) (string, error) {
|
||||
fileURL := urlWithVersionArch(installerType, targetVersion)
|
||||
|
||||
// Clean up temp directory on error
|
||||
var success bool
|
||||
defer func() {
|
||||
if !success {
|
||||
if err := os.RemoveAll(u.tempDir); err != nil {
|
||||
log.Errorf("error cleaning up temporary directory: %v", err)
|
||||
}
|
||||
}
|
||||
}()
|
||||
|
||||
fileName := path.Base(fileURL)
|
||||
if fileName == "." || fileName == "/" || fileName == "" {
|
||||
return "", fmt.Errorf("invalid file URL: %s", fileURL)
|
||||
}
|
||||
|
||||
outputFilePath := filepath.Join(u.tempDir, fileName)
|
||||
if err := downloader.DownloadToFile(ctx, downloader.DefaultRetryDelay, fileURL, outputFilePath); err != nil {
|
||||
return "", err
|
||||
}
|
||||
|
||||
success = true
|
||||
return outputFilePath, nil
|
||||
}
|
||||
|
||||
func (u *Installer) TempDir() string {
|
||||
return u.tempDir
|
||||
}
|
||||
|
||||
func (u *Installer) mkTempDir() error {
|
||||
if err := os.MkdirAll(u.tempDir, 0o755); err != nil {
|
||||
log.Debugf("failed to create tempdir: %s", u.tempDir)
|
||||
return err
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
func (u *Installer) copyUpdater() (string, error) {
|
||||
src, err := getServiceBinary()
|
||||
if err != nil {
|
||||
return "", fmt.Errorf("failed to get updater binary: %w", err)
|
||||
}
|
||||
|
||||
dst := filepath.Join(u.tempDir, updaterBinary)
|
||||
if err := copyFile(src, dst); err != nil {
|
||||
return "", fmt.Errorf("failed to copy updater binary: %w", err)
|
||||
}
|
||||
|
||||
if err := os.Chmod(dst, 0o755); err != nil {
|
||||
return "", fmt.Errorf("failed to set permissions: %w", err)
|
||||
}
|
||||
|
||||
return dst, nil
|
||||
}
|
||||
|
||||
func validateTargetVersion(targetVersion string) error {
|
||||
if targetVersion == "" {
|
||||
return fmt.Errorf("target version cannot be empty")
|
||||
}
|
||||
|
||||
_, err := goversion.NewVersion(targetVersion)
|
||||
if err != nil {
|
||||
return fmt.Errorf("invalid target version %q: %w", targetVersion, err)
|
||||
}
|
||||
|
||||
return nil
|
||||
}
|
||||
|
||||
func copyFile(src, dst string) error {
|
||||
log.Infof("copying %s to %s", src, dst)
|
||||
in, err := os.Open(src)
|
||||
if err != nil {
|
||||
return fmt.Errorf("open source: %w", err)
|
||||
}
|
||||
defer func() {
|
||||
if err := in.Close(); err != nil {
|
||||
log.Warnf("failed to close source file: %v", err)
|
||||
}
|
||||
}()
|
||||
|
||||
out, err := os.Create(dst)
|
||||
if err != nil {
|
||||
return fmt.Errorf("create destination: %w", err)
|
||||
}
|
||||
defer func() {
|
||||
if err := out.Close(); err != nil {
|
||||
log.Warnf("failed to close destination file: %v", err)
|
||||
}
|
||||
}()
|
||||
|
||||
if _, err := io.Copy(out, in); err != nil {
|
||||
return fmt.Errorf("copy: %w", err)
|
||||
}
|
||||
|
||||
return nil
|
||||
}
|
||||
|
||||
func getServiceDir() (string, error) {
|
||||
exePath, err := os.Executable()
|
||||
if err != nil {
|
||||
return "", err
|
||||
}
|
||||
return filepath.Dir(exePath), nil
|
||||
}
|
||||
|
||||
func getServiceBinary() (string, error) {
|
||||
return os.Executable()
|
||||
}
|
||||
|
||||
func isDryRunEnabled() bool {
|
||||
return strings.EqualFold(strings.TrimSpace(os.Getenv("NB_AUTO_UPDATE_DRY_RUN")), "true")
|
||||
}
|
||||
@@ -1,11 +0,0 @@
|
||||
package installer
|
||||
|
||||
import (
|
||||
"path/filepath"
|
||||
)
|
||||
|
||||
func (u *Installer) LogFiles() []string {
|
||||
return []string{
|
||||
filepath.Join(u.tempDir, LogFile),
|
||||
}
|
||||
}
|
||||
@@ -1,12 +0,0 @@
|
||||
package installer
|
||||
|
||||
import (
|
||||
"path/filepath"
|
||||
)
|
||||
|
||||
func (u *Installer) LogFiles() []string {
|
||||
return []string{
|
||||
filepath.Join(u.tempDir, msiLogFile),
|
||||
filepath.Join(u.tempDir, LogFile),
|
||||
}
|
||||
}
|
||||
@@ -1,238 +0,0 @@
|
||||
package installer
|
||||
|
||||
import (
|
||||
"context"
|
||||
"fmt"
|
||||
"os"
|
||||
"os/exec"
|
||||
"os/user"
|
||||
"path/filepath"
|
||||
"runtime"
|
||||
"strings"
|
||||
"syscall"
|
||||
"time"
|
||||
|
||||
log "github.com/sirupsen/logrus"
|
||||
)
|
||||
|
||||
const (
|
||||
daemonName = "netbird"
|
||||
updaterBinary = "updater"
|
||||
uiBinary = "/Applications/NetBird.app"
|
||||
|
||||
defaultTempDir = "/var/lib/netbird/tmp-install"
|
||||
|
||||
pkgDownloadURL = "https://github.com/netbirdio/netbird/releases/download/v%version/netbird_%version_darwin_%arch.pkg"
|
||||
)
|
||||
|
||||
var (
|
||||
binaryExtensions = []string{"pkg"}
|
||||
)
|
||||
|
||||
// Setup runs the installer with appropriate arguments and manages the daemon/UI state
|
||||
// This will be run by the updater process
|
||||
func (u *Installer) Setup(ctx context.Context, dryRun bool, installerFile string, daemonFolder string) (resultErr error) {
|
||||
resultHandler := NewResultHandler(u.tempDir)
|
||||
|
||||
// Always ensure daemon and UI are restarted after setup
|
||||
defer func() {
|
||||
log.Infof("write out result")
|
||||
var err error
|
||||
if resultErr == nil {
|
||||
err = resultHandler.WriteSuccess()
|
||||
} else {
|
||||
err = resultHandler.WriteErr(resultErr)
|
||||
}
|
||||
if err != nil {
|
||||
log.Errorf("failed to write update result: %v", err)
|
||||
}
|
||||
|
||||
// skip service restart if dry-run mode is enabled
|
||||
if dryRun {
|
||||
return
|
||||
}
|
||||
|
||||
log.Infof("starting daemon back")
|
||||
if err := u.startDaemon(daemonFolder); err != nil {
|
||||
log.Errorf("failed to start daemon: %v", err)
|
||||
}
|
||||
|
||||
log.Infof("starting UI back")
|
||||
if err := u.startUIAsUser(); err != nil {
|
||||
log.Errorf("failed to start UI: %v", err)
|
||||
}
|
||||
|
||||
}()
|
||||
|
||||
if dryRun {
|
||||
time.Sleep(7 * time.Second)
|
||||
log.Infof("dry-run mode enabled, skipping actual installation")
|
||||
resultErr = fmt.Errorf("dry-run mode enabled")
|
||||
return
|
||||
}
|
||||
|
||||
switch TypeOfInstaller(ctx) {
|
||||
case TypePKG:
|
||||
resultErr = u.installPkgFile(ctx, installerFile)
|
||||
case TypeHomebrew:
|
||||
resultErr = u.updateHomeBrew(ctx)
|
||||
}
|
||||
|
||||
return resultErr
|
||||
}
|
||||
|
||||
func (u *Installer) startDaemon(daemonFolder string) error {
|
||||
log.Infof("starting netbird service")
|
||||
ctx, cancel := context.WithTimeout(context.Background(), 15*time.Second)
|
||||
defer cancel()
|
||||
|
||||
cmd := exec.CommandContext(ctx, filepath.Join(daemonFolder, daemonName), "service", "start")
|
||||
if output, err := cmd.CombinedOutput(); err != nil {
|
||||
log.Warnf("failed to start netbird service: %v, output: %s", err, string(output))
|
||||
return err
|
||||
}
|
||||
log.Infof("netbird service started successfully")
|
||||
return nil
|
||||
}
|
||||
|
||||
func (u *Installer) startUIAsUser() error {
|
||||
log.Infof("starting netbird-ui: %s", uiBinary)
|
||||
|
||||
// Get the current console user
|
||||
cmd := exec.Command("stat", "-f", "%Su", "/dev/console")
|
||||
output, err := cmd.Output()
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to get console user: %w", err)
|
||||
}
|
||||
|
||||
username := strings.TrimSpace(string(output))
|
||||
if username == "" || username == "root" {
|
||||
return fmt.Errorf("no active user session found")
|
||||
}
|
||||
|
||||
log.Infof("starting UI for user: %s", username)
|
||||
|
||||
// Get user's UID
|
||||
userInfo, err := user.Lookup(username)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to lookup user %s: %w", username, err)
|
||||
}
|
||||
|
||||
// Start the UI process as the console user using launchctl
|
||||
// This ensures the app runs in the user's context with proper GUI access
|
||||
launchCmd := exec.Command("launchctl", "asuser", userInfo.Uid, "open", "-a", uiBinary)
|
||||
log.Infof("launchCmd: %s", launchCmd.String())
|
||||
// Set the user's home directory for proper macOS app behavior
|
||||
launchCmd.Env = append(os.Environ(), "HOME="+userInfo.HomeDir)
|
||||
log.Infof("set HOME environment variable: %s", userInfo.HomeDir)
|
||||
|
||||
if err := launchCmd.Start(); err != nil {
|
||||
return fmt.Errorf("failed to start UI process: %w", err)
|
||||
}
|
||||
|
||||
// Release the process so it can run independently
|
||||
if err := launchCmd.Process.Release(); err != nil {
|
||||
log.Warnf("failed to release UI process: %v", err)
|
||||
}
|
||||
|
||||
log.Infof("netbird-ui started successfully for user %s", username)
|
||||
return nil
|
||||
}
|
||||
|
||||
func (u *Installer) installPkgFile(ctx context.Context, path string) error {
|
||||
log.Infof("installing pkg file: %s", path)
|
||||
|
||||
// Kill any existing UI processes before installation
|
||||
// This ensures the postinstall script's "open $APP" will start the new version
|
||||
u.killUI()
|
||||
|
||||
volume := "/"
|
||||
|
||||
cmd := exec.CommandContext(ctx, "installer", "-pkg", path, "-target", volume)
|
||||
if err := cmd.Start(); err != nil {
|
||||
return fmt.Errorf("error running pkg file: %w", err)
|
||||
}
|
||||
log.Infof("installer started with PID %d", cmd.Process.Pid)
|
||||
if err := cmd.Wait(); err != nil {
|
||||
return fmt.Errorf("error running pkg file: %w", err)
|
||||
}
|
||||
log.Infof("pkg file installed successfully")
|
||||
return nil
|
||||
}
|
||||
|
||||
func (u *Installer) updateHomeBrew(ctx context.Context) error {
|
||||
log.Infof("updating homebrew")
|
||||
|
||||
// Kill any existing UI processes before upgrade
|
||||
// This ensures the new version will be started after upgrade
|
||||
u.killUI()
|
||||
|
||||
// Homebrew must be run as a non-root user
|
||||
// To find out which user installed NetBird using HomeBrew we can check the owner of our brew tap directory
|
||||
// Check both Apple Silicon and Intel Mac paths
|
||||
brewTapPath := "/opt/homebrew/Library/Taps/netbirdio/homebrew-tap/"
|
||||
brewBinPath := "/opt/homebrew/bin/brew"
|
||||
if _, err := os.Stat(brewTapPath); os.IsNotExist(err) {
|
||||
// Try Intel Mac path
|
||||
brewTapPath = "/usr/local/Homebrew/Library/Taps/netbirdio/homebrew-tap/"
|
||||
brewBinPath = "/usr/local/bin/brew"
|
||||
}
|
||||
|
||||
fileInfo, err := os.Stat(brewTapPath)
|
||||
if err != nil {
|
||||
return fmt.Errorf("error getting homebrew installation path info: %w", err)
|
||||
}
|
||||
|
||||
fileSysInfo, ok := fileInfo.Sys().(*syscall.Stat_t)
|
||||
if !ok {
|
||||
return fmt.Errorf("error checking file owner, sysInfo type is %T not *syscall.Stat_t", fileInfo.Sys())
|
||||
}
|
||||
|
||||
// Get username from UID
|
||||
brewUser, err := user.LookupId(fmt.Sprintf("%d", fileSysInfo.Uid))
|
||||
if err != nil {
|
||||
return fmt.Errorf("error looking up brew installer user: %w", err)
|
||||
}
|
||||
userName := brewUser.Username
|
||||
// Get user HOME, required for brew to run correctly
|
||||
// https://github.com/Homebrew/brew/issues/15833
|
||||
homeDir := brewUser.HomeDir
|
||||
|
||||
// Check if netbird-ui is installed (must run as the brew user, not root)
|
||||
checkUICmd := exec.CommandContext(ctx, "sudo", "-u", userName, brewBinPath, "list", "--formula", "netbirdio/tap/netbird-ui")
|
||||
checkUICmd.Env = append(os.Environ(), "HOME="+homeDir)
|
||||
uiInstalled := checkUICmd.Run() == nil
|
||||
|
||||
// Homebrew does not support installing specific versions
|
||||
// Thus it will always update to latest and ignore targetVersion
|
||||
upgradeArgs := []string{"-u", userName, brewBinPath, "upgrade", "netbirdio/tap/netbird"}
|
||||
if uiInstalled {
|
||||
upgradeArgs = append(upgradeArgs, "netbirdio/tap/netbird-ui")
|
||||
}
|
||||
|
||||
cmd := exec.CommandContext(ctx, "sudo", upgradeArgs...)
|
||||
cmd.Env = append(os.Environ(), "HOME="+homeDir)
|
||||
|
||||
if output, err := cmd.CombinedOutput(); err != nil {
|
||||
return fmt.Errorf("error running brew upgrade: %w, output: %s", err, string(output))
|
||||
}
|
||||
|
||||
log.Infof("homebrew updated successfully")
|
||||
return nil
|
||||
}
|
||||
|
||||
func (u *Installer) killUI() {
|
||||
log.Infof("killing existing netbird-ui processes")
|
||||
cmd := exec.Command("pkill", "-x", "netbird-ui")
|
||||
if output, err := cmd.CombinedOutput(); err != nil {
|
||||
// pkill returns exit code 1 if no processes matched, which is fine
|
||||
log.Debugf("pkill netbird-ui result: %v, output: %s", err, string(output))
|
||||
} else {
|
||||
log.Infof("netbird-ui processes killed")
|
||||
}
|
||||
}
|
||||
|
||||
func urlWithVersionArch(_ Type, version string) string {
|
||||
url := strings.ReplaceAll(pkgDownloadURL, "%version", version)
|
||||
return strings.ReplaceAll(url, "%arch", runtime.GOARCH)
|
||||
}
|
||||
@@ -1,213 +0,0 @@
|
||||
package installer
|
||||
|
||||
import (
|
||||
"context"
|
||||
"fmt"
|
||||
"os"
|
||||
"os/exec"
|
||||
"path/filepath"
|
||||
"runtime"
|
||||
"strings"
|
||||
"time"
|
||||
"unsafe"
|
||||
|
||||
log "github.com/sirupsen/logrus"
|
||||
"golang.org/x/sys/windows"
|
||||
)
|
||||
|
||||
const (
|
||||
daemonName = "netbird.exe"
|
||||
uiName = "netbird-ui.exe"
|
||||
updaterBinary = "updater.exe"
|
||||
|
||||
msiLogFile = "msi.log"
|
||||
|
||||
msiDownloadURL = "https://github.com/netbirdio/netbird/releases/download/v%version/netbird_installer_%version_windows_%arch.msi"
|
||||
exeDownloadURL = "https://github.com/netbirdio/netbird/releases/download/v%version/netbird_installer_%version_windows_%arch.exe"
|
||||
)
|
||||
|
||||
var (
|
||||
defaultTempDir = filepath.Join(os.Getenv("ProgramData"), "Netbird", "tmp-install")
|
||||
|
||||
// for the cleanup
|
||||
binaryExtensions = []string{"msi", "exe"}
|
||||
)
|
||||
|
||||
// Setup runs the installer with appropriate arguments and manages the daemon/UI state
|
||||
// This will be run by the updater process
|
||||
func (u *Installer) Setup(ctx context.Context, dryRun bool, installerFile string, daemonFolder string) (resultErr error) {
|
||||
resultHandler := NewResultHandler(u.tempDir)
|
||||
|
||||
// Always ensure daemon and UI are restarted after setup
|
||||
defer func() {
|
||||
log.Infof("starting daemon back")
|
||||
if err := u.startDaemon(daemonFolder); err != nil {
|
||||
log.Errorf("failed to start daemon: %v", err)
|
||||
}
|
||||
|
||||
log.Infof("starting UI back")
|
||||
if err := u.startUIAsUser(daemonFolder); err != nil {
|
||||
log.Errorf("failed to start UI: %v", err)
|
||||
}
|
||||
|
||||
log.Infof("write out result")
|
||||
var err error
|
||||
if resultErr == nil {
|
||||
err = resultHandler.WriteSuccess()
|
||||
} else {
|
||||
err = resultHandler.WriteErr(resultErr)
|
||||
}
|
||||
if err != nil {
|
||||
log.Errorf("failed to write update result: %v", err)
|
||||
}
|
||||
}()
|
||||
|
||||
if dryRun {
|
||||
log.Infof("dry-run mode enabled, skipping actual installation")
|
||||
resultErr = fmt.Errorf("dry-run mode enabled")
|
||||
return
|
||||
}
|
||||
|
||||
installerType, err := typeByFileExtension(installerFile)
|
||||
if err != nil {
|
||||
log.Debugf("%v", err)
|
||||
resultErr = err
|
||||
return
|
||||
}
|
||||
|
||||
var cmd *exec.Cmd
|
||||
switch installerType {
|
||||
case TypeExe:
|
||||
log.Infof("run exe installer: %s", installerFile)
|
||||
cmd = exec.CommandContext(ctx, installerFile, "/S")
|
||||
default:
|
||||
installerDir := filepath.Dir(installerFile)
|
||||
logPath := filepath.Join(installerDir, msiLogFile)
|
||||
log.Infof("run msi installer: %s", installerFile)
|
||||
cmd = exec.CommandContext(ctx, "msiexec.exe", "/i", filepath.Base(installerFile), "/quiet", "/qn", "/l*v", logPath)
|
||||
}
|
||||
|
||||
cmd.Dir = filepath.Dir(installerFile)
|
||||
|
||||
if resultErr = cmd.Start(); resultErr != nil {
|
||||
log.Errorf("error starting installer: %v", resultErr)
|
||||
return
|
||||
}
|
||||
|
||||
log.Infof("installer started with PID %d", cmd.Process.Pid)
|
||||
if resultErr = cmd.Wait(); resultErr != nil {
|
||||
log.Errorf("installer process finished with error: %v", resultErr)
|
||||
return
|
||||
}
|
||||
|
||||
return nil
|
||||
}
|
||||
|
||||
func (u *Installer) startDaemon(daemonFolder string) error {
|
||||
log.Infof("starting netbird service")
|
||||
ctx, cancel := context.WithTimeout(context.Background(), 15*time.Second)
|
||||
defer cancel()
|
||||
|
||||
cmd := exec.CommandContext(ctx, filepath.Join(daemonFolder, daemonName), "service", "start")
|
||||
if output, err := cmd.CombinedOutput(); err != nil {
|
||||
log.Debugf("failed to start netbird service: %v, output: %s", err, string(output))
|
||||
return err
|
||||
}
|
||||
log.Infof("netbird service started successfully")
|
||||
return nil
|
||||
}
|
||||
|
||||
func (u *Installer) startUIAsUser(daemonFolder string) error {
|
||||
uiPath := filepath.Join(daemonFolder, uiName)
|
||||
log.Infof("starting netbird-ui: %s", uiPath)
|
||||
|
||||
// Get the active console session ID
|
||||
sessionID := windows.WTSGetActiveConsoleSessionId()
|
||||
if sessionID == 0xFFFFFFFF {
|
||||
return fmt.Errorf("no active user session found")
|
||||
}
|
||||
|
||||
// Get the user token for that session
|
||||
var userToken windows.Token
|
||||
err := windows.WTSQueryUserToken(sessionID, &userToken)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to query user token: %w", err)
|
||||
}
|
||||
defer func() {
|
||||
if err := userToken.Close(); err != nil {
|
||||
log.Warnf("failed to close user token: %v", err)
|
||||
}
|
||||
}()
|
||||
|
||||
// Duplicate the token to a primary token
|
||||
var primaryToken windows.Token
|
||||
err = windows.DuplicateTokenEx(
|
||||
userToken,
|
||||
windows.MAXIMUM_ALLOWED,
|
||||
nil,
|
||||
windows.SecurityImpersonation,
|
||||
windows.TokenPrimary,
|
||||
&primaryToken,
|
||||
)
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to duplicate token: %w", err)
|
||||
}
|
||||
defer func() {
|
||||
if err := primaryToken.Close(); err != nil {
|
||||
log.Warnf("failed to close token: %v", err)
|
||||
}
|
||||
}()
|
||||
|
||||
// Prepare startup info
|
||||
var si windows.StartupInfo
|
||||
si.Cb = uint32(unsafe.Sizeof(si))
|
||||
si.Desktop = windows.StringToUTF16Ptr("winsta0\\default")
|
||||
|
||||
var pi windows.ProcessInformation
|
||||
|
||||
cmdLine, err := windows.UTF16PtrFromString(fmt.Sprintf("\"%s\"", uiPath))
|
||||
if err != nil {
|
||||
return fmt.Errorf("failed to convert path to UTF16: %w", err)
|
||||
}
|
||||
|
||||
creationFlags := uint32(0x00000200 | 0x00000008 | 0x00000400) // CREATE_NEW_PROCESS_GROUP | DETACHED_PROCESS | CREATE_UNICODE_ENVIRONMENT
|
||||
|
||||
err = windows.CreateProcessAsUser(
|
||||
primaryToken,
|
||||
nil,
|
||||
cmdLine,
|
||||
nil,
|
||||
nil,
|
||||
false,
|
||||
creationFlags,
|
||||
nil,
|
||||
nil,
|
||||
&si,
|
||||
&pi,
|
||||
)
|
||||
if err != nil {
|
||||
return fmt.Errorf("CreateProcessAsUser failed: %w", err)
|
||||
}
|
||||
|
||||
// Close handles
|
||||
if err := windows.CloseHandle(pi.Process); err != nil {
|
||||
log.Warnf("failed to close process handle: %v", err)
|
||||
}
|
||||
if err := windows.CloseHandle(pi.Thread); err != nil {
|
||||
log.Warnf("failed to close thread handle: %v", err)
|
||||
}
|
||||
|
||||
log.Infof("netbird-ui started successfully in session %d", sessionID)
|
||||
return nil
|
||||
}
|
||||
|
||||
func urlWithVersionArch(it Type, version string) string {
|
||||
var url string
|
||||
if it == TypeExe {
|
||||
url = exeDownloadURL
|
||||
} else {
|
||||
url = msiDownloadURL
|
||||
}
|
||||
url = strings.ReplaceAll(url, "%version", version)
|
||||
return strings.ReplaceAll(url, "%arch", runtime.GOARCH)
|
||||
}
|
||||
@@ -1,5 +0,0 @@
|
||||
package installer
|
||||
|
||||
const (
|
||||
LogFile = "installer.log"
|
||||
)
|
||||
@@ -1,15 +0,0 @@
|
||||
package installer
|
||||
|
||||
import (
|
||||
"os/exec"
|
||||
"syscall"
|
||||
)
|
||||
|
||||
// setUpdaterProcAttr configures the updater process to run in a new session,
|
||||
// making it independent of the parent daemon process. This ensures the updater
|
||||
// survives when the daemon is stopped during the pkg installation.
|
||||
func setUpdaterProcAttr(cmd *exec.Cmd) {
|
||||
cmd.SysProcAttr = &syscall.SysProcAttr{
|
||||
Setsid: true,
|
||||
}
|
||||
}
|
||||
Some files were not shown because too many files have changed in this diff Show More
Reference in New Issue
Block a user