Compare commits

...

75 Commits

Author SHA1 Message Date
snyk-bot
22df3afa78 fix: management/Dockerfile to reduce vulnerabilities
The following vulnerabilities are fixed with an upgrade:
- https://snyk.io/vuln/SNYK-UBUNTU2404-GLIBC-11031045
- https://snyk.io/vuln/SNYK-UBUNTU2404-PAM-11936905
- https://snyk.io/vuln/SNYK-UBUNTU2404-PAM-11936905
- https://snyk.io/vuln/SNYK-UBUNTU2404-PAM-11936905
- https://snyk.io/vuln/SNYK-UBUNTU2404-PAM-11936905
2025-09-30 03:47:02 +00:00
hakansa
17bab881f7 [client] Add Windows DNS Policies To GPO Path Always (#4460)
[client] Add Windows DNS Policies To GPO Path Always (#4460)
2025-09-26 16:42:18 +07:00
Vlad
25ed58328a [management] fix network map dns filter (#4547) 2025-09-25 16:29:14 +02:00
hakansa
644ed4b934 [client] Add WireGuard interface lifecycle monitoring (#4370)
* [client] Add WireGuard interface lifecycle monitoring
2025-09-25 15:36:26 +07:00
Pascal Fischer
58faa341d2 [management] Add logs for update channel (#4527) 2025-09-23 12:06:10 +02:00
Viktor Liu
5853b5553c [client] Skip interface for route lookup if it doesn't exist (#4524) 2025-09-22 14:32:00 +02:00
Zoltan Papp
998fb30e1e [client] Check the client status in the earlier phase (#4509)
This PR improves the NetBird client's status checking mechanism by implementing earlier detection of client state changes and better handling of connection lifecycle management. The key improvements focus on:

  • Enhanced status detection - Added waitForReady option to StatusRequest for improved client status handling
  • Better connection management - Improved context handling for signal and management gRPC connections• Reduced connection timeouts - Increased gRPC dial timeout from 3 to 10 seconds for better reliability
  • Cleaner error handling - Enhanced error propagation and context cancellation in retry loops

  Key Changes

  Core Status Improvements:
  - Added waitForReady optional field to StatusRequest proto (daemon.proto:190)
  - Enhanced status checking logic to detect client state changes earlier in the connection process
  - Improved handling of client permanent exit scenarios from retry loops

  Connection & Context Management:
  - Fixed context cancellation in management and signal client retry mechanisms
  - Added proper context propagation for Login operations
  - Enhanced gRPC connection handling with better timeout management

  Error Handling & Cleanup:
  - Moved feedback channels to upper layers for better separation of concerns
  - Improved error handling patterns throughout the client server implementation
  - Fixed synchronization issues and removed debug logging
2025-09-20 22:14:01 +02:00
Maycon Santos
e254b4cde5 [misc] Update SIGN_PIPE_VER to version 0.0.23 (#4521) 2025-09-20 10:24:04 +02:00
Zoltan Papp
ead1c618ba [client] Do not run up cmd if not needed in docker (#4508)
optimizes the NetBird client startup process by avoiding unnecessary login commands when the peer is already authenticated. The changes increase the default login timeout and expand the log message patterns used to detect successful authentication.

- Increased default login timeout from 1 to 5 seconds for more reliable authentication detection
- Enhanced log pattern matching to detect both registration and ready states
- Added extended regex support for more flexible pattern matching
2025-09-20 10:00:18 +02:00
Viktor Liu
55126f990c [client] Use native windows sock opts to avoid routing loops (#4314)
- Move `util/grpc` and `util/net` to `client` so `internal` packages can be accessed
 - Add methods to return the next best interface after the NetBird interface.
- Use `IP_UNICAST_IF` sock opt to force the outgoing interface for the NetBird `net.Dialer` and `net.ListenerConfig` to avoid routing loops. The interface is picked by the new route lookup method.
- Some refactoring to avoid import cycles
- Old behavior is available through `NB_USE_LEGACY_ROUTING=true` env var
2025-09-20 09:31:04 +02:00
Misha Bragin
90577682e4 Add a new product demo video (#4520) 2025-09-19 13:06:44 +02:00
Bethuel Mmbaga
dc30dcacce [management] Filter DNS records to include only peers to connect (#4517)
DNS record filtering to only include peers that a peer can connect to, reducing unnecessary DNS data in the peer's network map.

- Adds a new `filterZoneRecordsForPeers` function to filter DNS records based on peer connectivity
- Modifies `GetPeerNetworkMap` to use filtered DNS records instead of all records in the custom zone
- Includes comprehensive test coverage for the new filtering functionality
2025-09-18 18:57:07 +02:00
Diego Romar
2c87fa6236 [android] Add OnLoginSuccess callback to URLOpener interface (#4492)
The callback will be fired once login -> internal.Login
completes without errors
2025-09-18 15:07:42 +02:00
hakansa
ec8d83ade4 [client] [UI] Down & Up NetBird Async When Settings Updated
[client] [UI] Down & Up NetBird Async When Settings Updated
2025-09-18 18:13:29 +07:00
Bethuel Mmbaga
3130cce72d [management] Add rule ID validation for policy updates (#4499) 2025-09-15 21:08:16 +03:00
Zoltan Papp
bd23ab925e [client] Fix ICE latency handling (#4501)
The GetSelectedCandidatePair() does not carry the latency information.
2025-09-15 15:08:53 +02:00
Zoltan Papp
0c6f671a7c Refactor healthcheck sender and receiver to use configurable options (#4433) 2025-09-12 09:31:03 +02:00
Bethuel Mmbaga
cf7f6c355f [misc] Remove default zitadel admin user in deployment script (#4482)
* Delete default zitadel-admin user during initialization

Signed-off-by: bcmmbaga <bethuelmbaga12@gmail.com>

* Refactor

Signed-off-by: bcmmbaga <bethuelmbaga12@gmail.com>

---------

Signed-off-by: bcmmbaga <bethuelmbaga12@gmail.com>
2025-09-11 21:20:10 +02:00
Zoltan Papp
47e64d72db [client] Fix client status check (#4474)
The client status is not enough to protect the RPC calls from concurrency issues, because it is handled internally in the client in an asynchronous way.
2025-09-11 16:21:09 +02:00
Zoltan Papp
9e81e782e5 [client] Fix/v4 stun routing (#4430)
Deduplicate STUN package sending.
Originally, because every peer shared the same UDP address, the library could not distinguish which STUN message was associated with which candidate. As a result, the Pion library responded from all candidates for every STUN message.
2025-09-11 10:08:54 +02:00
Zoltan Papp
7aef0f67df [client] Implement environment variable handling for Android (#4440)
Some features can only be manipulated via environment variables. With this PR, environment variables can be managed from Android.
2025-09-08 18:42:42 +02:00
Maycon Santos
dba7ef667d [misc] Remove aur support and start service on ostree (#4461)
* Remove aur support and start service on ostree

The aur installation was adding many packages and installing more than just the client. For now is best to remove it and rely on binary install

Some users complained about ostree installation not starting the client, we add two explicit commands to it

* use  ${SUDO}

* fix if closure
2025-09-08 15:03:56 +02:00
Zoltan Papp
69d87343d2 [client] Debug information for connection (#4439)
Improve logging

Print the exact time when the first WireGuard handshake occurs
Print the steps for gathering system information
2025-09-08 14:51:34 +02:00
Bethuel Mmbaga
5113c70943 [management] Extends integration and peers manager (#4450) 2025-09-06 13:13:49 +03:00
Zoltan Papp
ad8fcda67b [client] Move some sys info to static place (#4446)
This PR refactors the system information collection code by moving static system information gathering to a dedicated location and separating platform-specific implementations. The primary goal is to improve code organization and maintainability by centralizing static info collection logic.

Key changes:
- Centralized static info collection into dedicated files with platform-specific implementations
- Moved `StaticInfo` struct definition to the main static_info.go file
- Added async initialization function `UpdateStaticInfoAsync()` across all platforms
2025-09-06 10:49:28 +02:00
Pascal Fischer
d33f88df82 [management] only allow user devices to be expired (#4445) 2025-09-05 18:11:23 +02:00
Zoltan Papp
786ca6fc79 Do not block Offer processing from relay worker (#4435)
- do not miss ICE offers when relay worker busy
- close p2p connection before recreate agent
2025-09-05 11:02:29 +02:00
Diego Romar
dfebdf1444 [internal] Add missing assignment of iFaceDiscover when netstack is disabled (#4444)
The internal updateInterfaces() function expects iFaceDiscover to not
be nil
2025-09-04 23:00:10 +02:00
Bethuel Mmbaga
a8dcff69c2 [management] Add peers manager to integrations (#4405) 2025-09-04 23:07:03 +03:00
Viktor Liu
71e944fa57 [relay] Let relay accept any origin (#4426) 2025-09-01 19:51:06 +02:00
Maycon Santos
d39fcfd62a [management] Add user approval (#4411)
This PR adds user approval functionality to the management system, allowing administrators to manually approve new users joining via domain matching. When enabled, users are blocked with pending approval status until explicitly approved by an admin.

Adds UserApprovalRequired setting to control manual user approval requirement
Introduces user approval and rejection endpoints with corresponding business logic
Prevents pending approval users from adding peers or logging in
2025-09-01 18:00:45 +02:00
Zoltan Papp
21368b38d9 [client] Update Pion ICE to the latest version (#4388)
- Update Pion version
- Update protobuf version
2025-09-01 10:42:01 +02:00
Maycon Santos
d817584f52 [misc] fix Windows client and management bench tests (#4424)
Windows tests had too many directories, causing issues to the payload via psexec.

Also migrated all checked benchmarks to send data to grafana.
2025-08-31 17:19:56 +02:00
Pascal Fischer
4d3dc3475d [management] remove duplicated removal of groups on peer delete (#4421) 2025-08-30 12:47:13 +02:00
Pascal Fischer
6fc50a438f [management] remove withContext from store methods (#4422) 2025-08-30 12:46:54 +02:00
Vlad
149559a06b [management] login filter to fix multiple peers connected with the same pub key (#3986) 2025-08-29 19:48:40 +02:00
Pascal Fischer
e14c6de203 [management] fix ephemeral flag on peer batch response (#4420) 2025-08-29 17:41:20 +02:00
Viktor Liu
d4c067f0af [client] Don't deactivate upstream resolvers on failure (#4128) 2025-08-29 17:40:05 +02:00
Pascal Fischer
dbefa8bd9f [management] remove lock and continue user update on failure (#4410) 2025-08-28 17:50:12 +02:00
Pascal Fischer
4fd10b9447 [management] split high latency grpc metrics (#4408) 2025-08-28 13:25:40 +02:00
Viktor Liu
aa595c3073 [client] Fix shared sock buffer allocation (#4409) 2025-08-28 13:25:16 +02:00
Vlad
99bd34c02a [signal] fix goroutines and memory leak on forward messages between peers (#3896) 2025-08-27 19:30:49 +03:00
Krzysztof Nazarewski (kdn)
7ce5507c05 [client] fix darwin dns always throwing err (#4403)
* fix: dns/host_darwin.go was missing if err != nil before throwing error
2025-08-27 09:59:39 +02:00
Pascal Fischer
0320bb7b35 [management] Report sync duration and login duration by accountID (#4406) 2025-08-26 22:32:12 +02:00
Viktor Liu
f063866ce8 [client] Add flag to configure MTU (#4213) 2025-08-26 16:00:14 +02:00
plusls
9f84165763 [client] Add netstack support for Android cli (#4319) 2025-08-26 15:40:01 +02:00
Pascal Fischer
3488a516c9 [management] Move increment network serial as last step of each transaction (#4397) 2025-08-25 17:27:07 +02:00
Pascal Fischer
5e273c121a [management] Remove store locks 3 (#4390) 2025-08-21 20:47:28 +02:00
Bethuel Mmbaga
968d95698e [management] Bump github.com/golang-jwt/jwt from 3.2.2+incompatible to 5.3.0 (#4375) 2025-08-21 15:02:51 +03:00
Pascal Fischer
28bef26537 [management] Remove Store Locks 2 (#4385) 2025-08-21 12:23:49 +02:00
Pascal Fischer
0d2845ea31 [management] optimize proxy network map (#4324) 2025-08-20 19:04:19 +02:00
Zoltan Papp
f425870c8e [client] Avoid duplicated agent close (#4383) 2025-08-20 18:50:51 +02:00
Pascal Fischer
f9d64a06c2 [management] Remove all store locks from grpc side (#4374) 2025-08-20 12:41:14 +02:00
hakansa
86555c44f7 refactor doc workflow (#4373)
refactor doc workflow (#4373)
2025-08-20 10:59:32 +03:00
Bastien Jeannelle
48792c64cd [misc] Fix confusing comment (#4376) 2025-08-20 00:12:00 +02:00
hakansa
533d93eb17 [management,client] Feat/exit node auto apply (#4272)
[management,client] Feat/exit node auto apply (#4272)
2025-08-19 18:19:24 +03:00
dependabot[bot]
9685411246 [misc] Bump golang.org/x/oauth2 from 0.24.0 to 0.27.0 (#4176)
Bumps [golang.org/x/oauth2](https://github.com/golang/oauth2) from 0.24.0 to 0.27.0
2025-08-19 16:26:46 +03:00
hakansa
d00a226556 [management] Add CreatedAt field to Peer and PeerBatch models (#4371)
[management] Add CreatedAt field to Peer and PeerBatch models (#4371)
2025-08-19 16:02:11 +03:00
Pascal Fischer
5d361b5421 [management] add nil handling for route domains (#4366) 2025-08-19 11:35:03 +02:00
dependabot[bot]
a889c4108b [misc] Bump github.com/containerd/containerd from 1.7.16 to 1.7.27 (#3527)
Bumps [github.com/containerd/containerd](https://github.com/containerd/containerd) from 1.7.16 to 1.7.27
2025-08-18 21:57:21 +03:00
Zoltan Papp
12cad854b2 [client] Fix/ice handshake (#4281)
In this PR, speed up the GRPC message processing, force the recreation of the ICE agent when getting a new, remote offer (do not wait for local STUN timeout).
2025-08-18 20:09:50 +02:00
Pascal Fischer
6a3846a8b7 [management] Remove save account calls (#4349) 2025-08-18 12:37:20 +02:00
Viktor Liu
7cd5dcae59 [client] Fix rule order for deny rules in peer ACLs (#4147) 2025-08-18 11:17:00 +02:00
Pascal Fischer
0e62325d46 [management] fail on geo location init failure (#4362) 2025-08-18 10:53:55 +02:00
Pascal Fischer
b3056d0937 [management] Use DI containers for server bootstrapping (#4343) 2025-08-15 17:14:48 +02:00
Zoltan Papp
ab853ac2a5 [server] Add MySQL initialization script and update Docker configuration (#4345) 2025-08-14 17:53:59 +02:00
Misha Bragin
e97f853909 Improve wording in the NetBird client app (#4316) 2025-08-13 22:03:48 +02:00
hakansa
70db8751d7 [client] Add --disable-update-settings flag to the service (#4335)
[client] Add --disable-update-settings flag to the service (#4335)
2025-08-13 21:05:12 +03:00
Zoltan Papp
86a00ab4af Fix Go tarball version in FreeBSD build configuration (#4339) 2025-08-13 13:52:11 +02:00
Zoltan Papp
3d4b502126 [server] Add health check HTTP endpoint for Relay server (#4297)
The health check endpoint listens on a dedicated HTTP server.
By default, it is available at 0.0.0.0:9000/health. This can be configured using the --health-listen-address flag.

The results are cached for 3 seconds to avoid excessive calls.

The health check performs the following:

Checks the number of active listeners.
Validates each listener via WebSocket and QUIC dials, including TLS certificate verification.
2025-08-13 10:40:04 +02:00
Bethuel Mmbaga
a4e8647aef [management] Enable flow groups (#4230)
Adds the ability to limit traffic events logging to specific peer groups
2025-08-13 00:00:40 +03:00
Viktor Liu
160b811e21 [client] Distinguish between NXDOMAIN and NODATA in the dns forwarder (#4321) 2025-08-12 15:59:42 +02:00
Viktor Liu
5e607cf4e9 [client] Skip dns upstream servers pointing to our dns server IP to prevent loops (#4330) 2025-08-12 15:41:23 +02:00
Viktor Liu
0fdb944058 [client] Create NRPT rules separately per domain (#4329) 2025-08-12 15:40:37 +02:00
Zoltan Papp
ccbabd9e2a Add pprof support for Relay server (#4203) 2025-08-12 12:24:24 +02:00
341 changed files with 11470 additions and 4378 deletions

View File

@@ -16,19 +16,29 @@ jobs:
steps:
- name: Read PR body
id: body
shell: bash
run: |
BODY=$(jq -r '.pull_request.body // ""' "$GITHUB_EVENT_PATH")
echo "body<<EOF" >> $GITHUB_OUTPUT
echo "$BODY" >> $GITHUB_OUTPUT
echo "EOF" >> $GITHUB_OUTPUT
set -euo pipefail
BODY_B64=$(jq -r '.pull_request.body // "" | @base64' "$GITHUB_EVENT_PATH")
{
echo "body_b64=$BODY_B64"
} >> "$GITHUB_OUTPUT"
- name: Validate checkbox selection
id: validate
shell: bash
env:
BODY_B64: ${{ steps.body.outputs.body_b64 }}
run: |
body='${{ steps.body.outputs.body }}'
set -euo pipefail
if ! body="$(printf '%s' "$BODY_B64" | base64 -d)"; then
echo "::error::Failed to decode PR body from base64. Data may be corrupted or missing."
exit 1
fi
added_checked=$(printf '%s' "$body" | grep -Ei '^[[:space:]]*-\s*\[x\]\s*I added/updated documentation' | wc -l | tr -d '[:space:]' || true)
noneed_checked=$(printf '%s' "$body" | grep -Ei '^[[:space:]]*-\s*\[x\]\s*Documentation is \*\*not needed\*\*' | wc -l | tr -d '[:space:]' || true)
added_checked=$(printf "%s" "$body" | grep -E '^- \[x\] I added/updated documentation' -i | wc -l | tr -d ' ')
noneed_checked=$(printf "%s" "$body" | grep -E '^- \[x\] Documentation is \*\*not needed\*\*' -i | wc -l | tr -d ' ')
if [ "$added_checked" -eq 1 ] && [ "$noneed_checked" -eq 1 ]; then
echo "::error::Choose exactly one: either 'docs added' OR 'not needed'."
@@ -41,30 +51,35 @@ jobs:
fi
if [ "$added_checked" -eq 1 ]; then
echo "mode=added" >> $GITHUB_OUTPUT
echo "mode=added" >> "$GITHUB_OUTPUT"
else
echo "mode=noneed" >> $GITHUB_OUTPUT
echo "mode=noneed" >> "$GITHUB_OUTPUT"
fi
- name: Extract docs PR URL (when 'docs added')
if: steps.validate.outputs.mode == 'added'
id: extract
shell: bash
env:
BODY_B64: ${{ steps.body.outputs.body_b64 }}
run: |
body='${{ steps.body.outputs.body }}'
set -euo pipefail
body="$(printf '%s' "$BODY_B64" | base64 -d)"
# Strictly require HTTPS and that it's a PR in netbirdio/docs
# Examples accepted:
# https://github.com/netbirdio/docs/pull/1234
url=$(printf "%s" "$body" | grep -Eo 'https://github\.com/netbirdio/docs/pull/[0-9]+' | head -n1 || true)
# e.g., https://github.com/netbirdio/docs/pull/1234
url="$(printf '%s' "$body" | grep -Eo 'https://github\.com/netbirdio/docs/pull/[0-9]+' | head -n1 || true)"
if [ -z "$url" ]; then
if [ -z "${url:-}" ]; then
echo "::error::You checked 'docs added' but didn't include a valid HTTPS PR link to netbirdio/docs (e.g., https://github.com/netbirdio/docs/pull/1234)."
exit 1
fi
pr_number=$(echo "$url" | sed -E 's#.*/pull/([0-9]+)$#\1#')
echo "url=$url" >> $GITHUB_OUTPUT
echo "pr_number=$pr_number" >> $GITHUB_OUTPUT
pr_number="$(printf '%s' "$url" | sed -E 's#.*/pull/([0-9]+)$#\1#')"
{
echo "url=$url"
echo "pr_number=$pr_number"
} >> "$GITHUB_OUTPUT"
- name: Verify docs PR exists (and is open or merged)
if: steps.validate.outputs.mode == 'added'

View File

@@ -25,8 +25,7 @@ jobs:
release: "14.2"
prepare: |
pkg install -y curl pkgconf xorg
LATEST_VERSION=$(curl -s https://go.dev/VERSION?m=text|head -n 1)
GO_TARBALL="$LATEST_VERSION.freebsd-amd64.tar.gz"
GO_TARBALL="go1.23.12.freebsd-amd64.tar.gz"
GO_URL="https://go.dev/dl/$GO_TARBALL"
curl -vLO "$GO_URL"
tar -C /usr/local -vxzf "$GO_TARBALL"

View File

@@ -217,7 +217,7 @@ jobs:
- arch: "386"
raceFlag: ""
- arch: "amd64"
raceFlag: ""
raceFlag: "-race"
runs-on: ubuntu-22.04
steps:
- name: Install Go
@@ -382,6 +382,32 @@ jobs:
store: [ 'sqlite', 'postgres' ]
runs-on: ubuntu-22.04
steps:
- name: Create Docker network
run: docker network create promnet
- name: Start Prometheus Pushgateway
run: docker run -d --name pushgateway --network promnet -p 9091:9091 prom/pushgateway
- name: Start Prometheus (for Pushgateway forwarding)
run: |
echo '
global:
scrape_interval: 15s
scrape_configs:
- job_name: "pushgateway"
static_configs:
- targets: ["pushgateway:9091"]
remote_write:
- url: ${{ secrets.GRAFANA_URL }}
basic_auth:
username: ${{ secrets.GRAFANA_USER }}
password: ${{ secrets.GRAFANA_API_KEY }}
' > prometheus.yml
docker run -d --name prometheus --network promnet \
-v $PWD/prometheus.yml:/etc/prometheus/prometheus.yml \
-p 9090:9090 \
prom/prometheus
- name: Install Go
uses: actions/setup-go@v5
with:
@@ -428,9 +454,10 @@ jobs:
CGO_ENABLED=1 GOARCH=${{ matrix.arch }} \
NETBIRD_STORE_ENGINE=${{ matrix.store }} \
CI=true \
GIT_BRANCH=${{ github.ref_name }} \
go test -tags devcert -run=^$ -bench=. \
-exec 'sudo --preserve-env=CI,NETBIRD_STORE_ENGINE' \
-timeout 20m ./management/... ./shared/management/...
-exec 'sudo --preserve-env=CI,NETBIRD_STORE_ENGINE,GIT_BRANCH,GITHUB_RUN_ID' \
-timeout 20m ./management/... ./shared/management/... $(go list ./management/... ./shared/management/... | grep -v -e /management/server/http)
api_benchmark:
name: "Management / Benchmark (API)"
@@ -521,7 +548,7 @@ jobs:
-run=^$ \
-bench=. \
-exec 'sudo --preserve-env=CI,NETBIRD_STORE_ENGINE,GIT_BRANCH,GITHUB_RUN_ID' \
-timeout 20m ./management/... ./shared/management/...
-timeout 20m ./management/server/http/...
api_integration_test:
name: "Management / Integration"
@@ -571,4 +598,4 @@ jobs:
CI=true \
go test -tags=integration \
-exec 'sudo --preserve-env=CI,NETBIRD_STORE_ENGINE' \
-timeout 20m ./management/... ./shared/management/...
-timeout 20m ./management/server/http/...

View File

@@ -63,7 +63,7 @@ jobs:
- run: PsExec64 -s -w ${{ github.workspace }} C:\hostedtoolcache\windows\go\${{ steps.go.outputs.go-version }}\x64\bin\go.exe env -w GOMODCACHE=${{ env.cache }}
- run: PsExec64 -s -w ${{ github.workspace }} C:\hostedtoolcache\windows\go\${{ steps.go.outputs.go-version }}\x64\bin\go.exe env -w GOCACHE=${{ env.modcache }}
- run: PsExec64 -s -w ${{ github.workspace }} C:\hostedtoolcache\windows\go\${{ steps.go.outputs.go-version }}\x64\bin\go.exe mod tidy
- run: echo "files=$(go list ./... | ForEach-Object { $_ } | Where-Object { $_ -notmatch '/management' })" >> $env:GITHUB_ENV
- run: echo "files=$(go list ./... | ForEach-Object { $_ } | Where-Object { $_ -notmatch '/management' } | Where-Object { $_ -notmatch '/relay' } | Where-Object { $_ -notmatch '/signal' })" >> $env:GITHUB_ENV
- name: test
run: PsExec64 -s -w ${{ github.workspace }} cmd.exe /c "C:\hostedtoolcache\windows\go\${{ steps.go.outputs.go-version }}\x64\bin\go.exe test -tags=devcert -timeout 10m -p 1 ${{ env.files }} > test-out.txt 2>&1"

View File

@@ -9,7 +9,7 @@ on:
pull_request:
env:
SIGN_PIPE_VER: "v0.0.22"
SIGN_PIPE_VER: "v0.0.23"
GORELEASER_VER: "v2.3.2"
PRODUCT_NAME: "NetBird"
COPYRIGHT: "NetBird GmbH"

View File

@@ -83,6 +83,15 @@ jobs:
- name: Checkout code
uses: actions/checkout@v4
- name: Setup MySQL privileges
if: matrix.store == 'mysql'
run: |
sleep 10
mysql -h 127.0.0.1 -u root -pmysqlroot -e "
GRANT SYSTEM_VARIABLES_ADMIN ON *.* TO 'netbird'@'%';
FLUSH PRIVILEGES;
"
- name: cp setup.env
run: cp infrastructure_files/tests/setup.env infrastructure_files/

View File

@@ -1,3 +1,4 @@
<div align="center">
<br/>
<br/>
@@ -52,7 +53,7 @@
### Open Source Network Security in a Single Platform
<img width="1188" alt="centralized-network-management 1" src="https://github.com/user-attachments/assets/c28cc8e4-15d2-4d2f-bb97-a6433db39d56" />
https://github.com/user-attachments/assets/10cec749-bb56-4ab3-97af-4e38850108d2
### NetBird on Lawrence Systems (Video)
[![Watch the video](https://img.youtube.com/vi/Kwrff6h0rEw/0.jpg)](https://www.youtube.com/watch?v=Kwrff6h0rEw)

View File

@@ -18,7 +18,7 @@ ENV \
NB_LOG_FILE="console,/var/log/netbird/client.log" \
NB_DAEMON_ADDR="unix:///var/run/netbird.sock" \
NB_ENTRYPOINT_SERVICE_TIMEOUT="5" \
NB_ENTRYPOINT_LOGIN_TIMEOUT="1"
NB_ENTRYPOINT_LOGIN_TIMEOUT="5"
ENTRYPOINT [ "/usr/local/bin/netbird-entrypoint.sh" ]

View File

@@ -4,6 +4,7 @@ package android
import (
"context"
"os"
"slices"
"sync"
@@ -18,7 +19,7 @@ import (
"github.com/netbirdio/netbird/client/internal/stdnet"
"github.com/netbirdio/netbird/client/system"
"github.com/netbirdio/netbird/formatter"
"github.com/netbirdio/netbird/util/net"
"github.com/netbirdio/netbird/client/net"
)
// ConnectionListener export internal Listener for mobile
@@ -83,7 +84,8 @@ func NewClient(cfgFile string, androidSDKVersion int, deviceName string, uiVersi
}
// Run start the internal client. It is a blocker function
func (c *Client) Run(urlOpener URLOpener, dns *DNSList, dnsReadyListener DnsReadyListener) error {
func (c *Client) Run(urlOpener URLOpener, dns *DNSList, dnsReadyListener DnsReadyListener, envList *EnvList) error {
exportEnvList(envList)
cfg, err := profilemanager.UpdateOrCreateConfig(profilemanager.ConfigInput{
ConfigPath: c.cfgFile,
})
@@ -118,7 +120,8 @@ func (c *Client) Run(urlOpener URLOpener, dns *DNSList, dnsReadyListener DnsRead
// 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(dns *DNSList, dnsReadyListener DnsReadyListener) error {
func (c *Client) RunWithoutLogin(dns *DNSList, dnsReadyListener DnsReadyListener, envList *EnvList) error {
exportEnvList(envList)
cfg, err := profilemanager.UpdateOrCreateConfig(profilemanager.ConfigInput{
ConfigPath: c.cfgFile,
})
@@ -249,3 +252,14 @@ func (c *Client) SetConnectionListener(listener ConnectionListener) {
func (c *Client) RemoveConnectionListener() {
c.recorder.RemoveConnectionListener()
}
func exportEnvList(list *EnvList) {
if list == nil {
return
}
for k, v := range list.AllItems() {
if err := os.Setenv(k, v); err != nil {
log.Errorf("could not set env variable %s: %v", k, err)
}
}
}

View File

@@ -0,0 +1,32 @@
package android
import "github.com/netbirdio/netbird/client/internal/peer"
var (
// EnvKeyNBForceRelay Exported for Android java client
EnvKeyNBForceRelay = peer.EnvKeyNBForceRelay
)
// EnvList wraps a Go map for export to Java
type EnvList struct {
data map[string]string
}
// NewEnvList creates a new EnvList
func NewEnvList() *EnvList {
return &EnvList{data: make(map[string]string)}
}
// Put adds a key-value pair
func (el *EnvList) Put(key, value string) {
el.data[key] = value
}
// Get retrieves a value by key
func (el *EnvList) Get(key string) string {
return el.data[key]
}
func (el *EnvList) AllItems() map[string]string {
return el.data
}

View File

@@ -33,6 +33,7 @@ type ErrListener interface {
// the backend want to show an url for the user
type URLOpener interface {
Open(string)
OnLoginSuccess()
}
// Auth can register or login new client
@@ -181,6 +182,11 @@ func (a *Auth) login(urlOpener URLOpener) error {
err = a.withBackOff(a.ctx, func() error {
err := internal.Login(a.ctx, a.config, "", jwtToken)
if err == nil {
go urlOpener.OnLoginSuccess()
}
if s, ok := gstatus.FromError(err); ok && (s.Code() == codes.InvalidArgument || s.Code() == codes.PermissionDenied) {
return nil
}

View File

@@ -33,7 +33,7 @@ var (
var debugCmd = &cobra.Command{
Use: "debug",
Short: "Debugging commands",
Long: "Provides commands for debugging and logging control within the NetBird daemon.",
Long: "Commands for debugging and logging within the NetBird daemon.",
}
var debugBundleCmd = &cobra.Command{

View File

@@ -14,7 +14,8 @@ import (
var downCmd = &cobra.Command{
Use: "down",
Short: "down netbird connections",
Short: "Disconnect from the NetBird network",
Long: "Disconnect the NetBird client from the network and management service. This will terminate all active connections with the remote peers.",
RunE: func(cmd *cobra.Command, args []string) error {
SetFlagsFromEnvVars(rootCmd)
@@ -26,7 +27,7 @@ var downCmd = &cobra.Command{
return err
}
ctx, cancel := context.WithTimeout(context.Background(), time.Second*7)
ctx, cancel := context.WithTimeout(context.Background(), time.Second*20)
defer cancel()
conn, err := DialClientGRPCServer(ctx, daemonAddr)

View File

@@ -31,7 +31,8 @@ func init() {
var loginCmd = &cobra.Command{
Use: "login",
Short: "login to the NetBird Management Service (first run)",
Short: "Log in to the NetBird network",
Long: "Log in to the NetBird network using a setup key or SSO",
RunE: func(cmd *cobra.Command, args []string) error {
if err := setEnvAndFlags(cmd); err != nil {
return fmt.Errorf("set env and flags: %v", err)
@@ -226,7 +227,7 @@ func doForegroundLogin(ctx context.Context, cmd *cobra.Command, setupKey string,
}
// update host's static platform and system information
system.UpdateStaticInfo()
system.UpdateStaticInfoAsync()
configFilePath, err := activeProf.FilePath()
if err != nil {

View File

@@ -14,7 +14,8 @@ import (
var logoutCmd = &cobra.Command{
Use: "deregister",
Aliases: []string{"logout"},
Short: "deregister from the NetBird Management Service and delete peer",
Short: "Deregister from the NetBird management service and delete this peer",
Long: "This command will deregister the current peer from the NetBird management service and all associated configuration. Use with caution as this will remove the peer from the network.",
RunE: func(cmd *cobra.Command, args []string) error {
SetFlagsFromEnvVars(rootCmd)

View File

@@ -15,7 +15,7 @@ var appendFlag bool
var networksCMD = &cobra.Command{
Use: "networks",
Aliases: []string{"routes"},
Short: "Manage networks",
Short: "Manage connections to NetBird Networks and Resources",
Long: `Commands to list, select, or deselect networks. Replaces the "routes" command.`,
}

View File

@@ -16,13 +16,13 @@ import (
var profileCmd = &cobra.Command{
Use: "profile",
Short: "manage NetBird profiles",
Long: `Manage NetBird profiles, allowing you to list, switch, and remove profiles.`,
Short: "Manage NetBird client profiles",
Long: `Commands to list, add, remove, and switch profiles. Profiles allow you to maintain different accounts in one client app.`,
}
var profileListCmd = &cobra.Command{
Use: "list",
Short: "list all profiles",
Short: "List all profiles",
Long: `List all available profiles in the NetBird client.`,
Aliases: []string{"ls"},
RunE: listProfilesFunc,
@@ -30,7 +30,7 @@ var profileListCmd = &cobra.Command{
var profileAddCmd = &cobra.Command{
Use: "add <profile_name>",
Short: "add a new profile",
Short: "Add a new profile",
Long: `Add a new profile to the NetBird client. The profile name must be unique.`,
Args: cobra.ExactArgs(1),
RunE: addProfileFunc,
@@ -38,16 +38,16 @@ var profileAddCmd = &cobra.Command{
var profileRemoveCmd = &cobra.Command{
Use: "remove <profile_name>",
Short: "remove a profile",
Long: `Remove a profile from the NetBird client. The profile must not be active.`,
Short: "Remove a profile",
Long: `Remove a profile from the NetBird client. The profile must not be inactive.`,
Args: cobra.ExactArgs(1),
RunE: removeProfileFunc,
}
var profileSelectCmd = &cobra.Command{
Use: "select <profile_name>",
Short: "select a profile",
Long: `Select a profile to be the active profile in the NetBird client. The profile must exist.`,
Short: "Select a profile",
Long: `Make the specified profile active. This will switch the client to use the selected profile's configuration.`,
Args: cobra.ExactArgs(1),
RunE: selectProfileFunc,
}

View File

@@ -39,6 +39,7 @@ const (
extraIFaceBlackListFlag = "extra-iface-blacklist"
dnsRouteIntervalFlag = "dns-router-interval"
enableLazyConnectionFlag = "enable-lazy-connection"
mtuFlag = "mtu"
)
var (
@@ -72,7 +73,9 @@ var (
anonymizeFlag bool
dnsRouteInterval time.Duration
lazyConnEnabled bool
mtu uint16
profilesDisabled bool
updateSettingsDisabled bool
rootCmd = &cobra.Command{
Use: "netbird",
@@ -228,7 +231,7 @@ func FlagNameToEnvVar(cmdFlag string, prefix string) string {
// DialClientGRPCServer returns client connection to the daemon server.
func DialClientGRPCServer(ctx context.Context, addr string) (*grpc.ClientConn, error) {
ctx, cancel := context.WithTimeout(ctx, time.Second*3)
ctx, cancel := context.WithTimeout(ctx, time.Second*10)
defer cancel()
return grpc.DialContext(

View File

@@ -54,6 +54,7 @@ func TestSetFlagsFromEnvVars(t *testing.T) {
cmd.PersistentFlags().StringVar(&interfaceName, interfaceNameFlag, iface.WgInterfaceDefault, "WireGuard interface name")
cmd.PersistentFlags().BoolVar(&rosenpassEnabled, enableRosenpassFlag, false, "Enable Rosenpass feature Rosenpass.")
cmd.PersistentFlags().Uint16Var(&wireguardPort, wireguardPortFlag, iface.DefaultWgPort, "WireGuard interface listening port")
cmd.PersistentFlags().Uint16Var(&mtu, mtuFlag, iface.DefaultMTU, "Set MTU (Maximum Transmission Unit) for the WireGuard interface")
t.Setenv("NB_EXTERNAL_IP_MAP", "abc,dec")
t.Setenv("NB_INTERFACE_NAME", "test-name")

View File

@@ -19,7 +19,7 @@ import (
var serviceCmd = &cobra.Command{
Use: "service",
Short: "manages NetBird service",
Short: "Manage the NetBird daemon service",
}
var (
@@ -42,7 +42,8 @@ func init() {
}
serviceCmd.AddCommand(runCmd, startCmd, stopCmd, restartCmd, svcStatusCmd, installCmd, uninstallCmd, reconfigureCmd)
serviceCmd.PersistentFlags().BoolVar(&profilesDisabled, "disable-profiles", false, "Disables profiles feature. If enabled, the client will not be able to change or edit any profile.")
serviceCmd.PersistentFlags().BoolVar(&profilesDisabled, "disable-profiles", false, "Disables profiles feature. If enabled, the client will not be able to change or edit any profile. To persist this setting, use: netbird service install --disable-profiles")
serviceCmd.PersistentFlags().BoolVar(&updateSettingsDisabled, "disable-update-settings", false, "Disables update settings feature. If enabled, the client will not be able to change or edit any settings. To persist this setting, use: netbird service install --disable-update-settings")
rootCmd.PersistentFlags().StringVarP(&serviceName, "service", "s", defaultServiceName, "Netbird system service name")
serviceEnvDesc := `Sets extra environment variables for the service. ` +

View File

@@ -27,7 +27,7 @@ func (p *program) Start(svc service.Service) error {
log.Info("starting NetBird service") //nolint
// Collect static system and platform information
system.UpdateStaticInfo()
system.UpdateStaticInfoAsync()
// in any case, even if configuration does not exists we run daemon to serve CLI gRPC API.
p.serv = grpc.NewServer()
@@ -61,7 +61,7 @@ func (p *program) Start(svc service.Service) error {
}
}
serverInstance := server.New(p.ctx, util.FindFirstLogPath(logFiles), configPath, profilesDisabled)
serverInstance := server.New(p.ctx, util.FindFirstLogPath(logFiles), configPath, profilesDisabled, updateSettingsDisabled)
if err := serverInstance.Start(); err != nil {
log.Fatalf("failed to start daemon: %v", err)
}

View File

@@ -49,6 +49,14 @@ func buildServiceArguments() []string {
args = append(args, "--log-file", logFile)
}
if profilesDisabled {
args = append(args, "--disable-profiles")
}
if updateSettingsDisabled {
args = append(args, "--disable-update-settings")
}
return args
}
@@ -99,7 +107,7 @@ func createServiceConfigForInstall() (*service.Config, error) {
var installCmd = &cobra.Command{
Use: "install",
Short: "installs NetBird service",
Short: "Install NetBird service",
RunE: func(cmd *cobra.Command, args []string) error {
if err := setupServiceCommand(cmd); err != nil {
return err

View File

@@ -40,7 +40,7 @@ var sshCmd = &cobra.Command{
return nil
},
Short: "connect to a remote SSH server",
Short: "Connect to a remote SSH server",
RunE: func(cmd *cobra.Command, args []string) error {
SetFlagsFromEnvVars(rootCmd)
SetFlagsFromEnvVars(cmd)

View File

@@ -32,7 +32,8 @@ var (
var statusCmd = &cobra.Command{
Use: "status",
Short: "status of the Netbird Service",
Short: "Display NetBird client status",
Long: "Display the current status of the NetBird client, including connection status, peer information, and network details.",
RunE: statusFunc,
}

View File

@@ -9,32 +9,31 @@ import (
"github.com/golang/mock/gomock"
"github.com/stretchr/testify/require"
"go.opentelemetry.io/otel"
"google.golang.org/grpc"
"github.com/netbirdio/management-integrations/integrations"
clientProto "github.com/netbirdio/netbird/client/proto"
client "github.com/netbirdio/netbird/client/server"
"github.com/netbirdio/netbird/management/internals/server/config"
mgmt "github.com/netbirdio/netbird/management/server"
"github.com/netbirdio/netbird/management/server/activity"
"github.com/netbirdio/netbird/management/server/groups"
"github.com/netbirdio/netbird/management/server/integrations/port_forwarding"
"github.com/netbirdio/netbird/management/server/peers"
"github.com/netbirdio/netbird/management/server/permissions"
"github.com/netbirdio/netbird/management/server/settings"
"github.com/netbirdio/netbird/management/server/store"
"github.com/netbirdio/netbird/management/server/telemetry"
"github.com/netbirdio/netbird/management/server/types"
"github.com/netbirdio/netbird/util"
"google.golang.org/grpc"
"github.com/netbirdio/management-integrations/integrations"
clientProto "github.com/netbirdio/netbird/client/proto"
client "github.com/netbirdio/netbird/client/server"
mgmtProto "github.com/netbirdio/netbird/shared/management/proto"
mgmt "github.com/netbirdio/netbird/management/server"
sigProto "github.com/netbirdio/netbird/shared/signal/proto"
sig "github.com/netbirdio/netbird/signal/server"
"github.com/netbirdio/netbird/util"
)
func startTestingServices(t *testing.T) string {
t.Helper()
config := &types.Config{}
config := &config.Config{}
_, err := util.ReadJson("../testdata/management.json", config)
if err != nil {
t.Fatal(err)
@@ -69,7 +68,7 @@ func startSignal(t *testing.T) (*grpc.Server, net.Listener) {
return s, lis
}
func startManagement(t *testing.T, config *types.Config, testFile string) (*grpc.Server, net.Listener) {
func startManagement(t *testing.T, config *config.Config, testFile string) (*grpc.Server, net.Listener) {
t.Helper()
lis, err := net.Listen("tcp", ":0")
@@ -88,15 +87,21 @@ func startManagement(t *testing.T, config *types.Config, testFile string) (*grpc
if err != nil {
return nil, nil
}
iv, _ := integrations.NewIntegratedValidator(context.Background(), eventStore)
metrics, err := telemetry.NewDefaultAppMetrics(context.Background())
require.NoError(t, err)
ctrl := gomock.NewController(t)
t.Cleanup(ctrl.Finish)
settingsMockManager := settings.NewMockManager(ctrl)
permissionsManagerMock := permissions.NewMockManager(ctrl)
peersmanager := peers.NewManager(store, permissionsManagerMock)
settingsManagerMock := settings.NewMockManager(ctrl)
iv, _ := integrations.NewIntegratedValidator(context.Background(), peersmanager, settingsManagerMock, eventStore)
metrics, err := telemetry.NewDefaultAppMetrics(context.Background())
require.NoError(t, err)
settingsMockManager := settings.NewMockManager(ctrl)
groupsManager := groups.NewManagerMock()
settingsMockManager.EXPECT().
GetSettings(gomock.Any(), gomock.Any(), gomock.Any()).
@@ -108,7 +113,7 @@ func startManagement(t *testing.T, config *types.Config, testFile string) (*grpc
t.Fatal(err)
}
secretsManager := mgmt.NewTimeBasedAuthSecretsManager(peersUpdateManager, config.TURNConfig, config.Relay, settingsMockManager)
secretsManager := mgmt.NewTimeBasedAuthSecretsManager(peersUpdateManager, config.TURNConfig, config.Relay, settingsMockManager, groupsManager)
mgmtServer, err := mgmt.NewServer(context.Background(), config, accountManager, settingsMockManager, peersUpdateManager, secretsManager, nil, nil, nil, &mgmt.MockIntegratedValidator{})
if err != nil {
t.Fatal(err)
@@ -134,7 +139,7 @@ func startClientDaemon(
s := grpc.NewServer()
server := client.New(ctx,
"", "", false)
"", "", false, false)
if err := server.Start(); err != nil {
t.Fatal(err)
}

View File

@@ -53,7 +53,8 @@ var (
upCmd = &cobra.Command{
Use: "up",
Short: "install, login and start NetBird client",
Short: "Connect to the NetBird network",
Long: "Connect to the NetBird network using the provided setup key or SSO auth. This command will bring up the WireGuard interface, connect to the management server, and establish peer-to-peer connections with other peers in the network if required.",
RunE: upFunc,
}
)
@@ -62,6 +63,7 @@ func init() {
upCmd.PersistentFlags().BoolVarP(&foregroundMode, "foreground-mode", "F", false, "start service in foreground")
upCmd.PersistentFlags().StringVar(&interfaceName, interfaceNameFlag, iface.WgInterfaceDefault, "WireGuard interface name")
upCmd.PersistentFlags().Uint16Var(&wireguardPort, wireguardPortFlag, iface.DefaultWgPort, "WireGuard interface listening port")
upCmd.PersistentFlags().Uint16Var(&mtu, mtuFlag, iface.DefaultMTU, "Set MTU (Maximum Transmission Unit) for the WireGuard interface")
upCmd.PersistentFlags().BoolVarP(&networkMonitor, networkMonitorFlag, "N", networkMonitor,
`Manage network monitoring. Defaults to true on Windows and macOS, false on Linux and FreeBSD. `+
`E.g. --network-monitor=false to disable or --network-monitor=true to enable.`,
@@ -228,7 +230,9 @@ func runInDaemonMode(ctx context.Context, cmd *cobra.Command, pm *profilemanager
client := proto.NewDaemonServiceClient(conn)
status, err := client.Status(ctx, &proto.StatusRequest{})
status, err := client.Status(ctx, &proto.StatusRequest{
WaitForReady: func() *bool { b := true; return &b }(),
})
if err != nil {
return fmt.Errorf("unable to get daemon status: %v", err)
}
@@ -356,6 +360,11 @@ func setupSetConfigReq(customDNSAddressConverted []byte, cmd *cobra.Command, pro
req.WireguardPort = &p
}
if cmd.Flag(mtuFlag).Changed {
m := int64(mtu)
req.Mtu = &m
}
if cmd.Flag(networkMonitorFlag).Changed {
req.NetworkMonitor = &networkMonitor
}
@@ -435,6 +444,13 @@ func setupConfig(customDNSAddressConverted []byte, cmd *cobra.Command, configFil
ic.WireguardPort = &p
}
if cmd.Flag(mtuFlag).Changed {
if err := iface.ValidateMTU(mtu); err != nil {
return nil, err
}
ic.MTU = &mtu
}
if cmd.Flag(networkMonitorFlag).Changed {
ic.NetworkMonitor = &networkMonitor
}
@@ -532,6 +548,14 @@ func setupLoginRequest(providedSetupKey string, customDNSAddressConverted []byte
loginRequest.WireguardPort = &wp
}
if cmd.Flag(mtuFlag).Changed {
if err := iface.ValidateMTU(mtu); err != nil {
return nil, err
}
m := int64(mtu)
loginRequest.Mtu = &m
}
if cmd.Flag(networkMonitorFlag).Changed {
loginRequest.NetworkMonitor = &networkMonitor
}

View File

@@ -9,7 +9,7 @@ import (
var (
versionCmd = &cobra.Command{
Use: "version",
Short: "prints NetBird version",
Short: "Print the NetBird's client application version",
Run: func(cmd *cobra.Command, args []string) {
cmd.SetOut(cmd.OutOrStdout())
cmd.Println(version.NetbirdVersion())

View File

@@ -135,7 +135,7 @@ func (c *Client) Start(startCtx context.Context) error {
// either startup error (permanent backoff err) or nil err (successful engine up)
// TODO: make after-startup backoff err available
run := make(chan struct{}, 1)
run := make(chan struct{})
clientErr := make(chan error, 1)
go func() {
if err := client.Run(run); err != nil {

View File

@@ -12,7 +12,7 @@ import (
firewall "github.com/netbirdio/netbird/client/firewall/manager"
"github.com/netbirdio/netbird/client/internal/statemanager"
nbnet "github.com/netbirdio/netbird/util/net"
nbnet "github.com/netbirdio/netbird/client/net"
)
const (
@@ -85,7 +85,7 @@ func (m *aclManager) AddPeerFiltering(
) ([]firewall.Rule, error) {
chain := chainNameInputRules
ipsetName = transformIPsetName(ipsetName, sPort, dPort)
ipsetName = transformIPsetName(ipsetName, sPort, dPort, action)
specs := filterRuleSpecs(ip, string(protocol), sPort, dPort, action, ipsetName)
mangleSpecs := slices.Clone(specs)
@@ -135,7 +135,14 @@ func (m *aclManager) AddPeerFiltering(
return nil, fmt.Errorf("rule already exists")
}
if err := m.iptablesClient.Append(tableFilter, chain, specs...); err != nil {
// Insert DROP rules at the beginning, append ACCEPT rules at the end
if action == firewall.ActionDrop {
// Insert at the beginning of the chain (position 1)
err = m.iptablesClient.Insert(tableFilter, chain, 1, specs...)
} else {
err = m.iptablesClient.Append(tableFilter, chain, specs...)
}
if err != nil {
return nil, err
}
@@ -388,17 +395,25 @@ func actionToStr(action firewall.Action) string {
return "DROP"
}
func transformIPsetName(ipsetName string, sPort, dPort *firewall.Port) string {
switch {
case ipsetName == "":
func transformIPsetName(ipsetName string, sPort, dPort *firewall.Port, action firewall.Action) string {
if ipsetName == "" {
return ""
}
// Include action in the ipset name to prevent squashing rules with different actions
actionSuffix := ""
if action == firewall.ActionDrop {
actionSuffix = "-drop"
}
switch {
case sPort != nil && dPort != nil:
return ipsetName + "-sport-dport"
return ipsetName + "-sport-dport" + actionSuffix
case sPort != nil:
return ipsetName + "-sport"
return ipsetName + "-sport" + actionSuffix
case dPort != nil:
return ipsetName + "-dport"
return ipsetName + "-dport" + actionSuffix
default:
return ipsetName
return ipsetName + actionSuffix
}
}

View File

@@ -3,6 +3,7 @@ package iptables
import (
"fmt"
"net/netip"
"strings"
"testing"
"time"
@@ -15,7 +16,7 @@ import (
var ifaceMock = &iFaceMock{
NameFunc: func() string {
return "lo"
return "wg-test"
},
AddressFunc: func() wgaddr.Address {
return wgaddr.Address{
@@ -109,10 +110,84 @@ func TestIptablesManager(t *testing.T) {
})
}
func TestIptablesManagerDenyRules(t *testing.T) {
ipv4Client, err := iptables.NewWithProtocol(iptables.ProtocolIPv4)
require.NoError(t, err)
manager, err := Create(ifaceMock)
require.NoError(t, err)
require.NoError(t, manager.Init(nil))
defer func() {
err := manager.Close(nil)
require.NoError(t, err)
}()
t.Run("add deny rule", func(t *testing.T) {
ip := netip.MustParseAddr("10.20.0.3")
port := &fw.Port{Values: []uint16{22}}
rule, err := manager.AddPeerFiltering(nil, ip.AsSlice(), "tcp", nil, port, fw.ActionDrop, "deny-ssh")
require.NoError(t, err, "failed to add deny rule")
require.NotEmpty(t, rule, "deny rule should not be empty")
// Verify the rule was added by checking iptables
for _, r := range rule {
rr := r.(*Rule)
checkRuleSpecs(t, ipv4Client, rr.chain, true, rr.specs...)
}
})
t.Run("deny rule precedence test", func(t *testing.T) {
ip := netip.MustParseAddr("10.20.0.4")
port := &fw.Port{Values: []uint16{80}}
// Add accept rule first
_, err := manager.AddPeerFiltering(nil, ip.AsSlice(), "tcp", nil, port, fw.ActionAccept, "accept-http")
require.NoError(t, err, "failed to add accept rule")
// Add deny rule second for same IP/port - this should take precedence
_, err = manager.AddPeerFiltering(nil, ip.AsSlice(), "tcp", nil, port, fw.ActionDrop, "deny-http")
require.NoError(t, err, "failed to add deny rule")
// Inspect the actual iptables rules to verify deny rule comes before accept rule
rules, err := ipv4Client.List("filter", chainNameInputRules)
require.NoError(t, err, "failed to list iptables rules")
// Debug: print all rules
t.Logf("All iptables rules in chain %s:", chainNameInputRules)
for i, rule := range rules {
t.Logf(" [%d] %s", i, rule)
}
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)
if strings.Contains(rule, "deny-http") && strings.Contains(rule, "80") {
denyRuleIndex = i
}
}
if strings.Contains(rule, "ACCEPT") {
t.Logf("Found ACCEPT rule at index %d: %s", i, rule)
if strings.Contains(rule, "accept-http") && strings.Contains(rule, "80") {
acceptRuleIndex = i
}
}
}
require.NotEqual(t, -1, denyRuleIndex, "deny rule should exist in iptables")
require.NotEqual(t, -1, acceptRuleIndex, "accept rule should exist in iptables")
require.Less(t, denyRuleIndex, acceptRuleIndex,
"deny rule should come before accept rule in iptables chain (deny at index %d, accept at index %d)",
denyRuleIndex, acceptRuleIndex)
})
}
func TestIptablesManagerIPSet(t *testing.T) {
mock := &iFaceMock{
NameFunc: func() string {
return "lo"
return "wg-test"
},
AddressFunc: func() wgaddr.Address {
return wgaddr.Address{
@@ -176,7 +251,7 @@ func checkRuleSpecs(t *testing.T, ipv4Client *iptables.IPTables, chainName strin
func TestIptablesCreatePerformance(t *testing.T) {
mock := &iFaceMock{
NameFunc: func() string {
return "lo"
return "wg-test"
},
AddressFunc: func() wgaddr.Address {
return wgaddr.Address{

View File

@@ -19,7 +19,7 @@ import (
"github.com/netbirdio/netbird/client/internal/routemanager/ipfwdstate"
"github.com/netbirdio/netbird/client/internal/routemanager/refcounter"
"github.com/netbirdio/netbird/client/internal/statemanager"
nbnet "github.com/netbirdio/netbird/util/net"
nbnet "github.com/netbirdio/netbird/client/net"
)
// constants needed to manage and create iptable rules

View File

@@ -14,7 +14,7 @@ import (
firewall "github.com/netbirdio/netbird/client/firewall/manager"
"github.com/netbirdio/netbird/client/firewall/test"
nbnet "github.com/netbirdio/netbird/util/net"
nbnet "github.com/netbirdio/netbird/client/net"
)
func isIptablesSupported() bool {

View File

@@ -16,7 +16,7 @@ import (
"golang.org/x/sys/unix"
firewall "github.com/netbirdio/netbird/client/firewall/manager"
nbnet "github.com/netbirdio/netbird/util/net"
nbnet "github.com/netbirdio/netbird/client/net"
)
const (
@@ -341,30 +341,38 @@ func (m *AclManager) addIOFiltering(
userData := []byte(ruleId)
chain := m.chainInputRules
nftRule := m.rConn.AddRule(&nftables.Rule{
rule := &nftables.Rule{
Table: m.workTable,
Chain: chain,
Exprs: mainExpressions,
UserData: userData,
})
}
// Insert DROP rules at the beginning, append ACCEPT rules at the end
var nftRule *nftables.Rule
if action == firewall.ActionDrop {
nftRule = m.rConn.InsertRule(rule)
} else {
nftRule = m.rConn.AddRule(rule)
}
if err := m.rConn.Flush(); err != nil {
return nil, fmt.Errorf(flushError, err)
}
rule := &Rule{
ruleStruct := &Rule{
nftRule: nftRule,
mangleRule: m.createPreroutingRule(expressions, userData),
nftSet: ipset,
ruleID: ruleId,
ip: ip,
}
m.rules[ruleId] = rule
m.rules[ruleId] = ruleStruct
if ipset != nil {
m.ipsetStore.AddReferenceToIpset(ipset.Name)
}
return rule, nil
return ruleStruct, nil
}
func (m *AclManager) createPreroutingRule(expressions []expr.Any, userData []byte) *nftables.Rule {

View File

@@ -2,6 +2,7 @@ package nftables
import (
"bytes"
"encoding/binary"
"fmt"
"net/netip"
"os/exec"
@@ -20,7 +21,7 @@ import (
var ifaceMock = &iFaceMock{
NameFunc: func() string {
return "lo"
return "wg-test"
},
AddressFunc: func() wgaddr.Address {
return wgaddr.Address{
@@ -103,9 +104,8 @@ func TestNftablesManager(t *testing.T) {
Kind: expr.VerdictAccept,
},
}
compareExprsIgnoringCounters(t, rules[0].Exprs, expectedExprs1)
expectedExprs2 := []expr.Any{
// Since DROP rules are inserted at position 0, the DROP rule comes first
expectedDropExprs := []expr.Any{
&expr.Payload{
DestRegister: 1,
Base: expr.PayloadBaseNetworkHeader,
@@ -141,7 +141,12 @@ func TestNftablesManager(t *testing.T) {
},
&expr.Verdict{Kind: expr.VerdictDrop},
}
require.ElementsMatch(t, rules[1].Exprs, expectedExprs2, "expected the same expressions")
// Compare DROP rule at position 0 (inserted first due to InsertRule)
compareExprsIgnoringCounters(t, rules[0].Exprs, expectedDropExprs)
// Compare connection tracking rule at position 1 (pushed down by DROP rule insertion)
compareExprsIgnoringCounters(t, rules[1].Exprs, expectedExprs1)
for _, r := range rule {
err = manager.DeletePeerRule(r)
@@ -160,10 +165,90 @@ func TestNftablesManager(t *testing.T) {
require.NoError(t, err, "failed to reset")
}
func TestNftablesManagerRuleOrder(t *testing.T) {
// This test verifies rule insertion order in nftables peer ACLs
// We add accept rule first, then deny rule to test ordering behavior
manager, err := Create(ifaceMock)
require.NoError(t, err)
require.NoError(t, manager.Init(nil))
defer func() {
err = manager.Close(nil)
require.NoError(t, err)
}()
ip := netip.MustParseAddr("100.96.0.2").Unmap()
testClient := &nftables.Conn{}
// Add accept rule first
_, err = manager.AddPeerFiltering(nil, ip.AsSlice(), fw.ProtocolTCP, nil, &fw.Port{Values: []uint16{80}}, fw.ActionAccept, "accept-http")
require.NoError(t, err, "failed to add accept rule")
// Add deny rule second for the same traffic
_, err = manager.AddPeerFiltering(nil, ip.AsSlice(), fw.ProtocolTCP, nil, &fw.Port{Values: []uint16{80}}, fw.ActionDrop, "deny-http")
require.NoError(t, err, "failed to add deny rule")
err = manager.Flush()
require.NoError(t, err, "failed to flush")
rules, err := testClient.GetRules(manager.aclManager.workTable, manager.aclManager.chainInputRules)
require.NoError(t, err, "failed to get rules")
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 int = -1, -1
for i, rule := range rules {
hasAcceptHTTPSet := false
hasDenyHTTPSet := false
hasPort80 := false
var action string
for _, e := range rule.Exprs {
// Check for set lookup
if lookup, ok := e.(*expr.Lookup); ok {
if lookup.SetName == "accept-http" {
hasAcceptHTTPSet = true
} else if lookup.SetName == "deny-http" {
hasDenyHTTPSet = true
}
}
// Check for port 80
if cmp, ok := e.(*expr.Cmp); ok {
if cmp.Op == expr.CmpOpEq && len(cmp.Data) == 2 && binary.BigEndian.Uint16(cmp.Data) == 80 {
hasPort80 = true
}
}
// Check for verdict
if verdict, ok := e.(*expr.Verdict); ok {
if verdict.Kind == expr.VerdictAccept {
action = "ACCEPT"
} else if verdict.Kind == expr.VerdictDrop {
action = "DROP"
}
}
}
if hasAcceptHTTPSet && hasPort80 && action == "ACCEPT" {
t.Logf("Rule [%d]: accept-http set + Port 80 + ACCEPT", i)
acceptRuleIndex = i
} else if hasDenyHTTPSet && hasPort80 && action == "DROP" {
t.Logf("Rule [%d]: deny-http set + Port 80 + DROP", i)
denyRuleIndex = i
}
}
require.NotEqual(t, -1, acceptRuleIndex, "accept rule should exist in nftables")
require.NotEqual(t, -1, denyRuleIndex, "deny rule should exist in nftables")
require.Less(t, denyRuleIndex, acceptRuleIndex,
"deny rule should come before accept rule in nftables chain (deny at index %d, accept at index %d)",
denyRuleIndex, acceptRuleIndex)
}
func TestNFtablesCreatePerformance(t *testing.T) {
mock := &iFaceMock{
NameFunc: func() string {
return "lo"
return "wg-test"
},
AddressFunc: func() wgaddr.Address {
return wgaddr.Address{

View File

@@ -22,7 +22,7 @@ import (
nbid "github.com/netbirdio/netbird/client/internal/acl/id"
"github.com/netbirdio/netbird/client/internal/routemanager/ipfwdstate"
"github.com/netbirdio/netbird/client/internal/routemanager/refcounter"
nbnet "github.com/netbirdio/netbird/util/net"
nbnet "github.com/netbirdio/netbird/client/net"
)
const (

View File

@@ -18,6 +18,7 @@ func (m *Manager) Close(stateManager *statemanager.Manager) error {
defer m.mutex.Unlock()
m.outgoingRules = make(map[netip.Addr]RuleSet)
m.incomingDenyRules = make(map[netip.Addr]RuleSet)
m.incomingRules = make(map[netip.Addr]RuleSet)
if m.udpTracker != nil {

View File

@@ -27,6 +27,7 @@ func (m *Manager) Close(*statemanager.Manager) error {
defer m.mutex.Unlock()
m.outgoingRules = make(map[netip.Addr]RuleSet)
m.incomingDenyRules = make(map[netip.Addr]RuleSet)
m.incomingRules = make(map[netip.Addr]RuleSet)
if m.udpTracker != nil {

View File

@@ -70,14 +70,13 @@ func (r RouteRules) Sort() {
// Manager userspace firewall manager
type Manager struct {
// outgoingRules is used for hooks only
outgoingRules map[netip.Addr]RuleSet
// incomingRules is used for filtering and hooks
incomingRules map[netip.Addr]RuleSet
routeRules RouteRules
decoders sync.Pool
wgIface common.IFaceMapper
nativeFirewall firewall.Manager
outgoingRules map[netip.Addr]RuleSet
incomingDenyRules map[netip.Addr]RuleSet
incomingRules map[netip.Addr]RuleSet
routeRules RouteRules
decoders sync.Pool
wgIface common.IFaceMapper
nativeFirewall firewall.Manager
mutex sync.RWMutex
@@ -186,6 +185,7 @@ func create(iface common.IFaceMapper, nativeFirewall firewall.Manager, disableSe
},
nativeFirewall: nativeFirewall,
outgoingRules: make(map[netip.Addr]RuleSet),
incomingDenyRules: make(map[netip.Addr]RuleSet),
incomingRules: make(map[netip.Addr]RuleSet),
wgIface: iface,
localipmanager: newLocalIPManager(),
@@ -417,10 +417,17 @@ func (m *Manager) AddPeerFiltering(
}
m.mutex.Lock()
if _, ok := m.incomingRules[r.ip]; !ok {
m.incomingRules[r.ip] = make(RuleSet)
var targetMap map[netip.Addr]RuleSet
if r.drop {
targetMap = m.incomingDenyRules
} else {
targetMap = m.incomingRules
}
m.incomingRules[r.ip][r.id] = r
if _, ok := targetMap[r.ip]; !ok {
targetMap[r.ip] = make(RuleSet)
}
targetMap[r.ip][r.id] = r
m.mutex.Unlock()
return []firewall.Rule{&r}, nil
}
@@ -507,10 +514,24 @@ func (m *Manager) DeletePeerRule(rule firewall.Rule) error {
return fmt.Errorf("delete rule: invalid rule type: %T", rule)
}
if _, ok := m.incomingRules[r.ip][r.id]; !ok {
var sourceMap map[netip.Addr]RuleSet
if r.drop {
sourceMap = m.incomingDenyRules
} else {
sourceMap = m.incomingRules
}
if ruleset, ok := sourceMap[r.ip]; ok {
if _, exists := ruleset[r.id]; !exists {
return fmt.Errorf("delete rule: no rule with such id: %v", r.id)
}
delete(ruleset, r.id)
if len(ruleset) == 0 {
delete(sourceMap, r.ip)
}
} else {
return fmt.Errorf("delete rule: no rule with such id: %v", r.id)
}
delete(m.incomingRules[r.ip], r.id)
return nil
}
@@ -572,7 +593,7 @@ func (m *Manager) UpdateSet(set firewall.Set, prefixes []netip.Prefix) error {
return nil
}
// FilterOutBound filters outgoing packets
// FilterOutbound filters outgoing packets
func (m *Manager) FilterOutbound(packetData []byte, size int) bool {
return m.filterOutbound(packetData, size)
}
@@ -761,7 +782,7 @@ func (m *Manager) filterInbound(packetData []byte, size int) bool {
// handleLocalTraffic handles local traffic.
// If it returns true, the packet should be dropped.
func (m *Manager) handleLocalTraffic(d *decoder, srcIP, dstIP netip.Addr, packetData []byte, size int) bool {
ruleID, blocked := m.peerACLsBlock(srcIP, packetData, m.incomingRules, d)
ruleID, blocked := m.peerACLsBlock(srcIP, d, packetData)
if blocked {
_, pnum := getProtocolFromPacket(d)
srcPort, dstPort := getPortsFromPacket(d)
@@ -971,26 +992,28 @@ func (m *Manager) isSpecialICMP(d *decoder) bool {
icmpType == layers.ICMPv4TypeTimeExceeded
}
func (m *Manager) peerACLsBlock(srcIP netip.Addr, packetData []byte, rules map[netip.Addr]RuleSet, d *decoder) ([]byte, bool) {
func (m *Manager) peerACLsBlock(srcIP netip.Addr, d *decoder, packetData []byte) ([]byte, bool) {
m.mutex.RLock()
defer m.mutex.RUnlock()
if m.isSpecialICMP(d) {
return nil, false
}
if mgmtId, filter, ok := validateRule(srcIP, packetData, rules[srcIP], d); ok {
if mgmtId, filter, ok := validateRule(srcIP, packetData, m.incomingDenyRules[srcIP], d); ok {
return mgmtId, filter
}
if mgmtId, filter, ok := validateRule(srcIP, packetData, rules[netip.IPv4Unspecified()], d); ok {
if mgmtId, filter, ok := validateRule(srcIP, packetData, m.incomingRules[srcIP], d); ok {
return mgmtId, filter
}
if mgmtId, filter, ok := validateRule(srcIP, packetData, m.incomingRules[netip.IPv4Unspecified()], d); ok {
return mgmtId, filter
}
if mgmtId, filter, ok := validateRule(srcIP, packetData, m.incomingRules[netip.IPv6Unspecified()], d); ok {
return mgmtId, filter
}
if mgmtId, filter, ok := validateRule(srcIP, packetData, rules[netip.IPv6Unspecified()], d); ok {
return mgmtId, filter
}
// Default policy: DROP ALL
return nil, true
}
@@ -1013,6 +1036,7 @@ func portsMatch(rulePort *firewall.Port, packetPort uint16) bool {
func validateRule(ip netip.Addr, packetData []byte, rules map[string]PeerRule, d *decoder) ([]byte, bool, bool) {
payloadLayer := d.decoded[1]
for _, rule := range rules {
if rule.matchByIP && ip.Compare(rule.ip) != 0 {
continue
@@ -1045,6 +1069,7 @@ func validateRule(ip netip.Addr, packetData []byte, rules map[string]PeerRule, d
return rule.mgmtId, rule.drop, true
}
}
return nil, false, false
}
@@ -1116,6 +1141,7 @@ func (m *Manager) AddUDPPacketHook(in bool, ip netip.Addr, dPort uint16, hook fu
m.mutex.Lock()
if in {
// Incoming UDP hooks are stored in allow rules map
if _, ok := m.incomingRules[r.ip]; !ok {
m.incomingRules[r.ip] = make(map[string]PeerRule)
}
@@ -1136,6 +1162,7 @@ func (m *Manager) RemovePacketHook(hookID string) error {
m.mutex.Lock()
defer m.mutex.Unlock()
// Check incoming hooks (stored in allow rules)
for _, arr := range m.incomingRules {
for _, r := range arr {
if r.id == hookID {
@@ -1144,6 +1171,7 @@ func (m *Manager) RemovePacketHook(hookID string) error {
}
}
}
// Check outgoing hooks
for _, arr := range m.outgoingRules {
for _, r := range arr {
if r.id == hookID {

View File

@@ -458,6 +458,31 @@ func TestPeerACLFiltering(t *testing.T) {
ruleAction: fw.ActionDrop,
shouldBeBlocked: true,
},
{
name: "Peer ACL - Drop rule should override accept all rule",
srcIP: "100.10.0.1",
dstIP: "100.10.0.100",
proto: fw.ProtocolTCP,
srcPort: 12345,
dstPort: 22,
ruleIP: "100.10.0.1",
ruleProto: fw.ProtocolTCP,
ruleDstPort: &fw.Port{Values: []uint16{22}},
ruleAction: fw.ActionDrop,
shouldBeBlocked: true,
},
{
name: "Peer ACL - Drop all traffic from specific IP",
srcIP: "100.10.0.99",
dstIP: "100.10.0.100",
proto: fw.ProtocolTCP,
srcPort: 12345,
dstPort: 80,
ruleIP: "100.10.0.99",
ruleProto: fw.ProtocolALL,
ruleAction: fw.ActionDrop,
shouldBeBlocked: true,
},
}
t.Run("Implicit DROP (no rules)", func(t *testing.T) {
@@ -468,13 +493,11 @@ func TestPeerACLFiltering(t *testing.T) {
for _, tc := range testCases {
t.Run(tc.name, func(t *testing.T) {
if tc.ruleAction == fw.ActionDrop {
// add general accept rule to test drop rule
// TODO: this only works because 0.0.0.0 is tested last, we need to implement order
// add general accept rule for the same IP to test drop rule precedence
rules, err := manager.AddPeerFiltering(
nil,
net.ParseIP("0.0.0.0"),
net.ParseIP(tc.ruleIP),
fw.ProtocolALL,
nil,
nil,

View File

@@ -136,9 +136,22 @@ func TestManagerDeleteRule(t *testing.T) {
return
}
// Check rules exist in appropriate maps
for _, r := range rule2 {
if _, ok := m.incomingRules[ip][r.ID()]; !ok {
t.Errorf("rule2 is not in the incomingRules")
peerRule, ok := r.(*PeerRule)
if !ok {
t.Errorf("rule should be a PeerRule")
continue
}
// Check if rule exists in deny or allow maps based on action
var found bool
if peerRule.drop {
_, found = m.incomingDenyRules[ip][r.ID()]
} else {
_, found = m.incomingRules[ip][r.ID()]
}
if !found {
t.Errorf("rule2 is not in the expected rules map")
}
}
@@ -150,9 +163,22 @@ func TestManagerDeleteRule(t *testing.T) {
}
}
// Check rules are removed from appropriate maps
for _, r := range rule2 {
if _, ok := m.incomingRules[ip][r.ID()]; ok {
t.Errorf("rule2 is not in the incomingRules")
peerRule, ok := r.(*PeerRule)
if !ok {
t.Errorf("rule should be a PeerRule")
continue
}
// Check if rule is removed from deny or allow maps based on action
var found bool
if peerRule.drop {
_, found = m.incomingDenyRules[ip][r.ID()]
} else {
_, found = m.incomingRules[ip][r.ID()]
}
if found {
t.Errorf("rule2 should be removed from the rules map")
}
}
}
@@ -196,16 +222,17 @@ func TestAddUDPPacketHook(t *testing.T) {
var addedRule PeerRule
if tt.in {
// Incoming UDP hooks are stored in allow rules map
if len(manager.incomingRules[tt.ip]) != 1 {
t.Errorf("expected 1 incoming rule, got %d", len(manager.incomingRules))
t.Errorf("expected 1 incoming rule, got %d", len(manager.incomingRules[tt.ip]))
return
}
for _, rule := range manager.incomingRules[tt.ip] {
addedRule = rule
}
} else {
if len(manager.outgoingRules) != 1 {
t.Errorf("expected 1 outgoing rule, got %d", len(manager.outgoingRules))
if len(manager.outgoingRules[tt.ip]) != 1 {
t.Errorf("expected 1 outgoing rule, got %d", len(manager.outgoingRules[tt.ip]))
return
}
for _, rule := range manager.outgoingRules[tt.ip] {
@@ -261,8 +288,8 @@ func TestManagerReset(t *testing.T) {
return
}
if len(m.outgoingRules) != 0 || len(m.incomingRules) != 0 {
t.Errorf("rules is not empty")
if len(m.outgoingRules) != 0 || len(m.incomingRules) != 0 || len(m.incomingDenyRules) != 0 {
t.Errorf("rules are not empty")
}
}

View File

@@ -314,7 +314,7 @@ func (m *Manager) buildConntrackStateMessage(d *decoder) string {
func (m *Manager) handleLocalDelivery(trace *PacketTrace, packetData []byte, d *decoder, srcIP, dstIP netip.Addr) bool {
trace.AddResult(StageRouting, "Packet destined for local delivery", true)
ruleId, blocked := m.peerACLsBlock(srcIP, packetData, m.incomingRules, d)
ruleId, blocked := m.peerACLsBlock(srcIP, d, packetData)
strRuleId := "<no id>"
if ruleId != nil {

View File

@@ -20,8 +20,9 @@ import (
"google.golang.org/grpc/credentials/insecure"
"google.golang.org/grpc/keepalive"
nbnet "github.com/netbirdio/netbird/client/net"
"github.com/netbirdio/netbird/util/embeddedroots"
nbnet "github.com/netbirdio/netbird/util/net"
)
func WithCustomDialer() grpc.DialOption {
@@ -57,7 +58,7 @@ func Backoff(ctx context.Context) backoff.BackOff {
return backoff.WithContext(b, ctx)
}
func CreateConnection(addr string, tlsEnabled bool) (*grpc.ClientConn, error) {
func CreateConnection(ctx context.Context, addr string, tlsEnabled bool) (*grpc.ClientConn, error) {
transportOption := grpc.WithTransportCredentials(insecure.NewCredentials())
if tlsEnabled {
certPool, err := x509.SystemCertPool()
@@ -71,7 +72,7 @@ func CreateConnection(addr string, tlsEnabled bool) (*grpc.ClientConn, error) {
}))
}
connCtx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
connCtx, cancel := context.WithTimeout(ctx, 30*time.Second)
defer cancel()
conn, err := grpc.DialContext(

View File

@@ -3,7 +3,7 @@ package bind
import (
wireguard "golang.zx2c4.com/wireguard/conn"
nbnet "github.com/netbirdio/netbird/util/net"
nbnet "github.com/netbirdio/netbird/client/net"
)
// TODO: This is most likely obsolete since the control fns should be called by the wrapped udpconn (ice_bind.go)

View File

@@ -8,15 +8,16 @@ import (
"runtime"
"sync"
"github.com/pion/stun/v2"
"github.com/pion/stun/v3"
"github.com/pion/transport/v3"
log "github.com/sirupsen/logrus"
"golang.org/x/net/ipv4"
"golang.org/x/net/ipv6"
wgConn "golang.zx2c4.com/wireguard/conn"
"github.com/netbirdio/netbird/client/iface/udpmux"
"github.com/netbirdio/netbird/client/iface/wgaddr"
nbnet "github.com/netbirdio/netbird/util/net"
nbnet "github.com/netbirdio/netbird/client/net"
)
type RecvMessage struct {
@@ -44,7 +45,7 @@ type ICEBind struct {
RecvChan chan RecvMessage
transportNet transport.Net
filterFn FilterFn
filterFn udpmux.FilterFn
endpoints map[netip.Addr]net.Conn
endpointsMu sync.Mutex
// every time when Close() is called (i.e. BindUpdate()) we need to close exit from the receiveRelayed and create a
@@ -54,12 +55,13 @@ type ICEBind struct {
closed bool
muUDPMux sync.Mutex
udpMux *UniversalUDPMuxDefault
udpMux *udpmux.UniversalUDPMuxDefault
address wgaddr.Address
mtu uint16
activityRecorder *ActivityRecorder
}
func NewICEBind(transportNet transport.Net, filterFn FilterFn, address wgaddr.Address) *ICEBind {
func NewICEBind(transportNet transport.Net, filterFn udpmux.FilterFn, address wgaddr.Address, mtu uint16) *ICEBind {
b, _ := wgConn.NewStdNetBind().(*wgConn.StdNetBind)
ib := &ICEBind{
StdNetBind: b,
@@ -69,6 +71,7 @@ func NewICEBind(transportNet transport.Net, filterFn FilterFn, address wgaddr.Ad
endpoints: make(map[netip.Addr]net.Conn),
closedChan: make(chan struct{}),
closed: true,
mtu: mtu,
address: address,
activityRecorder: NewActivityRecorder(),
}
@@ -80,6 +83,10 @@ func NewICEBind(transportNet transport.Net, filterFn FilterFn, address wgaddr.Ad
return ib
}
func (s *ICEBind) MTU() uint16 {
return s.mtu
}
func (s *ICEBind) Open(uport uint16) ([]wgConn.ReceiveFunc, uint16, error) {
s.closed = false
s.closedChanMu.Lock()
@@ -109,7 +116,7 @@ func (s *ICEBind) ActivityRecorder() *ActivityRecorder {
}
// GetICEMux returns the ICE UDPMux that was created and used by ICEBind
func (s *ICEBind) GetICEMux() (*UniversalUDPMuxDefault, error) {
func (s *ICEBind) GetICEMux() (*udpmux.UniversalUDPMuxDefault, error) {
s.muUDPMux.Lock()
defer s.muUDPMux.Unlock()
if s.udpMux == nil {
@@ -152,12 +159,13 @@ func (s *ICEBind) createIPv4ReceiverFn(pc *ipv4.PacketConn, conn *net.UDPConn, r
s.muUDPMux.Lock()
defer s.muUDPMux.Unlock()
s.udpMux = NewUniversalUDPMuxDefault(
UniversalUDPMuxParams{
s.udpMux = udpmux.NewUniversalUDPMuxDefault(
udpmux.UniversalUDPMuxParams{
UDPConn: nbnet.WrapPacketConn(conn),
Net: s.transportNet,
FilterFn: s.filterFn,
WGAddress: s.address,
MTU: s.mtu,
},
)
return func(bufs [][]byte, sizes []int, eps []wgConn.Endpoint) (n int, err error) {

View File

@@ -1,7 +0,0 @@
//go:build ios
package bind
func (m *UDPMuxDefault) notifyAddressRemoval(addr string) {
// iOS doesn't support nbnet hooks, so this is a no-op
}

View File

@@ -0,0 +1,9 @@
package bufsize
const (
// WGBufferOverhead represents the additional buffer space needed beyond MTU
// for WireGuard packet encapsulation (WG header + UDP + IP + safety margin)
// Original hardcoded buffers were 1500, default MTU is 1280, so overhead = 220
// TODO: Calculate this properly based on actual protocol overhead instead of using hardcoded difference
WGBufferOverhead = 220
)

View File

@@ -17,8 +17,8 @@ import (
"golang.zx2c4.com/wireguard/wgctrl/wgtypes"
"github.com/netbirdio/netbird/client/iface/bind"
nbnet "github.com/netbirdio/netbird/client/net"
"github.com/netbirdio/netbird/monotime"
nbnet "github.com/netbirdio/netbird/util/net"
)
const (
@@ -394,6 +394,13 @@ func toLastHandshake(stringVar string) (time.Time, error) {
if err != nil {
return time.Time{}, fmt.Errorf("parse handshake sec: %w", err)
}
// If sec is 0 (Unix epoch), return zero time instead
// This indicates no handshake has occurred
if sec == 0 {
return time.Time{}, nil
}
return time.Unix(sec, 0), nil
}
@@ -402,7 +409,7 @@ func toBytes(s string) (int64, error) {
}
func getFwmark() int {
if nbnet.AdvancedRouting() {
if nbnet.AdvancedRouting() && runtime.GOOS == "linux" {
return nbnet.ControlPlaneMark
}
return 0

View File

@@ -7,16 +7,17 @@ import (
wgdevice "golang.zx2c4.com/wireguard/device"
"github.com/netbirdio/netbird/client/iface/bind"
"github.com/netbirdio/netbird/client/iface/device"
"github.com/netbirdio/netbird/client/iface/udpmux"
"github.com/netbirdio/netbird/client/iface/wgaddr"
)
type WGTunDevice interface {
Create() (device.WGConfigurer, error)
Up() (*bind.UniversalUDPMuxDefault, error)
Up() (*udpmux.UniversalUDPMuxDefault, error)
UpdateAddr(address wgaddr.Address) error
WgAddress() wgaddr.Address
MTU() uint16
DeviceName() string
Close() error
FilteredDevice() *device.FilteredDevice

View File

@@ -13,6 +13,7 @@ import (
"github.com/netbirdio/netbird/client/iface/bind"
"github.com/netbirdio/netbird/client/iface/configurer"
"github.com/netbirdio/netbird/client/iface/udpmux"
"github.com/netbirdio/netbird/client/iface/wgaddr"
)
@@ -21,7 +22,7 @@ type WGTunDevice struct {
address wgaddr.Address
port int
key string
mtu int
mtu uint16
iceBind *bind.ICEBind
tunAdapter TunAdapter
disableDNS bool
@@ -29,11 +30,11 @@ type WGTunDevice struct {
name string
device *device.Device
filteredDevice *FilteredDevice
udpMux *bind.UniversalUDPMuxDefault
udpMux *udpmux.UniversalUDPMuxDefault
configurer WGConfigurer
}
func NewTunDevice(address wgaddr.Address, port int, key string, mtu int, iceBind *bind.ICEBind, tunAdapter TunAdapter, disableDNS bool) *WGTunDevice {
func NewTunDevice(address wgaddr.Address, port int, key string, mtu uint16, iceBind *bind.ICEBind, tunAdapter TunAdapter, disableDNS bool) *WGTunDevice {
return &WGTunDevice{
address: address,
port: port,
@@ -58,7 +59,7 @@ func (t *WGTunDevice) Create(routes []string, dns string, searchDomains []string
searchDomainsToString = ""
}
fd, err := t.tunAdapter.ConfigureInterface(t.address.String(), t.mtu, dns, searchDomainsToString, routesString)
fd, err := t.tunAdapter.ConfigureInterface(t.address.String(), int(t.mtu), dns, searchDomainsToString, routesString)
if err != nil {
log.Errorf("failed to create Android interface: %s", err)
return nil, err
@@ -88,7 +89,7 @@ func (t *WGTunDevice) Create(routes []string, dns string, searchDomains []string
}
return t.configurer, nil
}
func (t *WGTunDevice) Up() (*bind.UniversalUDPMuxDefault, error) {
func (t *WGTunDevice) Up() (*udpmux.UniversalUDPMuxDefault, error) {
err := t.device.Up()
if err != nil {
return nil, err
@@ -137,6 +138,10 @@ func (t *WGTunDevice) WgAddress() wgaddr.Address {
return t.address
}
func (t *WGTunDevice) MTU() uint16 {
return t.mtu
}
func (t *WGTunDevice) FilteredDevice() *FilteredDevice {
return t.filteredDevice
}

View File

@@ -13,6 +13,7 @@ import (
"github.com/netbirdio/netbird/client/iface/bind"
"github.com/netbirdio/netbird/client/iface/configurer"
"github.com/netbirdio/netbird/client/iface/udpmux"
"github.com/netbirdio/netbird/client/iface/wgaddr"
)
@@ -21,16 +22,16 @@ type TunDevice struct {
address wgaddr.Address
port int
key string
mtu int
mtu uint16
iceBind *bind.ICEBind
device *device.Device
filteredDevice *FilteredDevice
udpMux *bind.UniversalUDPMuxDefault
udpMux *udpmux.UniversalUDPMuxDefault
configurer WGConfigurer
}
func NewTunDevice(name string, address wgaddr.Address, port int, key string, mtu int, iceBind *bind.ICEBind) *TunDevice {
func NewTunDevice(name string, address wgaddr.Address, port int, key string, mtu uint16, iceBind *bind.ICEBind) *TunDevice {
return &TunDevice{
name: name,
address: address,
@@ -42,7 +43,7 @@ func NewTunDevice(name string, address wgaddr.Address, port int, key string, mtu
}
func (t *TunDevice) Create() (WGConfigurer, error) {
tunDevice, err := tun.CreateTUN(t.name, t.mtu)
tunDevice, err := tun.CreateTUN(t.name, int(t.mtu))
if err != nil {
return nil, fmt.Errorf("error creating tun device: %s", err)
}
@@ -71,7 +72,7 @@ func (t *TunDevice) Create() (WGConfigurer, error) {
return t.configurer, nil
}
func (t *TunDevice) Up() (*bind.UniversalUDPMuxDefault, error) {
func (t *TunDevice) Up() (*udpmux.UniversalUDPMuxDefault, error) {
err := t.device.Up()
if err != nil {
return nil, err
@@ -111,6 +112,10 @@ func (t *TunDevice) WgAddress() wgaddr.Address {
return t.address
}
func (t *TunDevice) MTU() uint16 {
return t.mtu
}
func (t *TunDevice) DeviceName() string {
return t.name
}

View File

@@ -14,6 +14,7 @@ import (
"github.com/netbirdio/netbird/client/iface/bind"
"github.com/netbirdio/netbird/client/iface/configurer"
"github.com/netbirdio/netbird/client/iface/udpmux"
"github.com/netbirdio/netbird/client/iface/wgaddr"
)
@@ -22,21 +23,23 @@ type TunDevice struct {
address wgaddr.Address
port int
key string
mtu uint16
iceBind *bind.ICEBind
tunFd int
device *device.Device
filteredDevice *FilteredDevice
udpMux *bind.UniversalUDPMuxDefault
udpMux *udpmux.UniversalUDPMuxDefault
configurer WGConfigurer
}
func NewTunDevice(name string, address wgaddr.Address, port int, key string, iceBind *bind.ICEBind, tunFd int) *TunDevice {
func NewTunDevice(name string, address wgaddr.Address, port int, key string, mtu uint16, iceBind *bind.ICEBind, tunFd int) *TunDevice {
return &TunDevice{
name: name,
address: address,
port: port,
key: key,
mtu: mtu,
iceBind: iceBind,
tunFd: tunFd,
}
@@ -81,7 +84,7 @@ func (t *TunDevice) Create() (WGConfigurer, error) {
return t.configurer, nil
}
func (t *TunDevice) Up() (*bind.UniversalUDPMuxDefault, error) {
func (t *TunDevice) Up() (*udpmux.UniversalUDPMuxDefault, error) {
err := t.device.Up()
if err != nil {
return nil, err
@@ -125,6 +128,10 @@ func (t *TunDevice) WgAddress() wgaddr.Address {
return t.address
}
func (t *TunDevice) MTU() uint16 {
return t.mtu
}
func (t *TunDevice) UpdateAddr(_ wgaddr.Address) error {
// todo implement
return nil

View File

@@ -12,11 +12,11 @@ import (
"golang.zx2c4.com/wireguard/device"
"golang.zx2c4.com/wireguard/tun/netstack"
"github.com/netbirdio/netbird/client/iface/bind"
"github.com/netbirdio/netbird/client/iface/configurer"
"github.com/netbirdio/netbird/client/iface/udpmux"
"github.com/netbirdio/netbird/client/iface/wgaddr"
nbnet "github.com/netbirdio/netbird/client/net"
"github.com/netbirdio/netbird/sharedsock"
nbnet "github.com/netbirdio/netbird/util/net"
)
type TunKernelDevice struct {
@@ -24,19 +24,19 @@ type TunKernelDevice struct {
address wgaddr.Address
wgPort int
key string
mtu int
mtu uint16
ctx context.Context
ctxCancel context.CancelFunc
transportNet transport.Net
link *wgLink
udpMuxConn net.PacketConn
udpMux *bind.UniversalUDPMuxDefault
udpMux *udpmux.UniversalUDPMuxDefault
filterFn bind.FilterFn
filterFn udpmux.FilterFn
}
func NewKernelDevice(name string, address wgaddr.Address, wgPort int, key string, mtu int, transportNet transport.Net) *TunKernelDevice {
func NewKernelDevice(name string, address wgaddr.Address, wgPort int, key string, mtu uint16, transportNet transport.Net) *TunKernelDevice {
ctx, cancel := context.WithCancel(context.Background())
return &TunKernelDevice{
ctx: ctx,
@@ -66,7 +66,7 @@ func (t *TunKernelDevice) Create() (WGConfigurer, error) {
// TODO: do a MTU discovery
log.Debugf("setting MTU: %d interface: %s", t.mtu, t.name)
if err := link.setMTU(t.mtu); err != nil {
if err := link.setMTU(int(t.mtu)); err != nil {
return nil, fmt.Errorf("set mtu: %w", err)
}
@@ -79,7 +79,7 @@ func (t *TunKernelDevice) Create() (WGConfigurer, error) {
return configurer, nil
}
func (t *TunKernelDevice) Up() (*bind.UniversalUDPMuxDefault, error) {
func (t *TunKernelDevice) Up() (*udpmux.UniversalUDPMuxDefault, error) {
if t.udpMux != nil {
return t.udpMux, nil
}
@@ -96,23 +96,19 @@ func (t *TunKernelDevice) Up() (*bind.UniversalUDPMuxDefault, error) {
return nil, err
}
rawSock, err := sharedsock.Listen(t.wgPort, sharedsock.NewIncomingSTUNFilter())
rawSock, err := sharedsock.Listen(t.wgPort, sharedsock.NewIncomingSTUNFilter(), t.mtu)
if err != nil {
return nil, err
}
var udpConn net.PacketConn = rawSock
if !nbnet.AdvancedRouting() {
udpConn = nbnet.WrapPacketConn(rawSock)
}
bindParams := bind.UniversalUDPMuxParams{
UDPConn: udpConn,
bindParams := udpmux.UniversalUDPMuxParams{
UDPConn: nbnet.WrapPacketConn(rawSock),
Net: t.transportNet,
FilterFn: t.filterFn,
WGAddress: t.address,
MTU: t.mtu,
}
mux := bind.NewUniversalUDPMuxDefault(bindParams)
mux := udpmux.NewUniversalUDPMuxDefault(bindParams)
go mux.ReadFromConn(t.ctx)
t.udpMuxConn = rawSock
t.udpMux = mux
@@ -158,6 +154,10 @@ func (t *TunKernelDevice) WgAddress() wgaddr.Address {
return t.address
}
func (t *TunKernelDevice) MTU() uint16 {
return t.mtu
}
func (t *TunKernelDevice) DeviceName() string {
return t.name
}

View File

@@ -1,6 +1,3 @@
//go:build !android
// +build !android
package device
import (
@@ -13,8 +10,9 @@ import (
"github.com/netbirdio/netbird/client/iface/bind"
"github.com/netbirdio/netbird/client/iface/configurer"
nbnetstack "github.com/netbirdio/netbird/client/iface/netstack"
"github.com/netbirdio/netbird/client/iface/udpmux"
"github.com/netbirdio/netbird/client/iface/wgaddr"
nbnet "github.com/netbirdio/netbird/util/net"
nbnet "github.com/netbirdio/netbird/client/net"
)
type TunNetstackDevice struct {
@@ -22,20 +20,20 @@ type TunNetstackDevice struct {
address wgaddr.Address
port int
key string
mtu int
mtu uint16
listenAddress string
iceBind *bind.ICEBind
device *device.Device
filteredDevice *FilteredDevice
nsTun *nbnetstack.NetStackTun
udpMux *bind.UniversalUDPMuxDefault
udpMux *udpmux.UniversalUDPMuxDefault
configurer WGConfigurer
net *netstack.Net
}
func NewNetstackDevice(name string, address wgaddr.Address, wgPort int, key string, mtu int, iceBind *bind.ICEBind, listenAddress string) *TunNetstackDevice {
func NewNetstackDevice(name string, address wgaddr.Address, wgPort int, key string, mtu uint16, iceBind *bind.ICEBind, listenAddress string) *TunNetstackDevice {
return &TunNetstackDevice{
name: name,
address: address,
@@ -47,7 +45,7 @@ func NewNetstackDevice(name string, address wgaddr.Address, wgPort int, key stri
}
}
func (t *TunNetstackDevice) Create() (WGConfigurer, error) {
func (t *TunNetstackDevice) create() (WGConfigurer, error) {
log.Info("create nbnetstack tun interface")
// TODO: get from service listener runtime IP
@@ -57,7 +55,7 @@ func (t *TunNetstackDevice) Create() (WGConfigurer, error) {
}
log.Debugf("netstack using address: %s", t.address.IP)
t.nsTun = nbnetstack.NewNetStackTun(t.listenAddress, t.address.IP, dnsAddr, t.mtu)
t.nsTun = nbnetstack.NewNetStackTun(t.listenAddress, t.address.IP, dnsAddr, int(t.mtu))
log.Debugf("netstack using dns address: %s", dnsAddr)
tunIface, net, err := t.nsTun.Create()
if err != nil {
@@ -83,7 +81,7 @@ func (t *TunNetstackDevice) Create() (WGConfigurer, error) {
return t.configurer, nil
}
func (t *TunNetstackDevice) Up() (*bind.UniversalUDPMuxDefault, error) {
func (t *TunNetstackDevice) Up() (*udpmux.UniversalUDPMuxDefault, error) {
if t.device == nil {
return nil, fmt.Errorf("device is not ready yet")
}
@@ -125,6 +123,10 @@ func (t *TunNetstackDevice) WgAddress() wgaddr.Address {
return t.address
}
func (t *TunNetstackDevice) MTU() uint16 {
return t.mtu
}
func (t *TunNetstackDevice) DeviceName() string {
return t.name
}

View File

@@ -0,0 +1,7 @@
//go:build android
package device
func (t *TunNetstackDevice) Create(routes []string, dns string, searchDomains []string) (WGConfigurer, error) {
return t.create()
}

View File

@@ -0,0 +1,7 @@
//go:build !android
package device
func (t *TunNetstackDevice) Create() (WGConfigurer, error) {
return t.create()
}

View File

@@ -12,6 +12,7 @@ import (
"github.com/netbirdio/netbird/client/iface/bind"
"github.com/netbirdio/netbird/client/iface/configurer"
"github.com/netbirdio/netbird/client/iface/udpmux"
"github.com/netbirdio/netbird/client/iface/wgaddr"
)
@@ -20,16 +21,16 @@ type USPDevice struct {
address wgaddr.Address
port int
key string
mtu int
mtu uint16
iceBind *bind.ICEBind
device *device.Device
filteredDevice *FilteredDevice
udpMux *bind.UniversalUDPMuxDefault
udpMux *udpmux.UniversalUDPMuxDefault
configurer WGConfigurer
}
func NewUSPDevice(name string, address wgaddr.Address, port int, key string, mtu int, iceBind *bind.ICEBind) *USPDevice {
func NewUSPDevice(name string, address wgaddr.Address, port int, key string, mtu uint16, iceBind *bind.ICEBind) *USPDevice {
log.Infof("using userspace bind mode")
return &USPDevice{
@@ -44,9 +45,9 @@ func NewUSPDevice(name string, address wgaddr.Address, port int, key string, mtu
func (t *USPDevice) Create() (WGConfigurer, error) {
log.Info("create tun interface")
tunIface, err := tun.CreateTUN(t.name, t.mtu)
tunIface, err := tun.CreateTUN(t.name, int(t.mtu))
if err != nil {
log.Debugf("failed to create tun interface (%s, %d): %s", t.name, t.mtu, err)
log.Debugf("failed to create tun interface (%s, %d): %s", t.name, int(t.mtu), err)
return nil, fmt.Errorf("error creating tun device: %s", err)
}
t.filteredDevice = newDeviceFilter(tunIface)
@@ -74,7 +75,7 @@ func (t *USPDevice) Create() (WGConfigurer, error) {
return t.configurer, nil
}
func (t *USPDevice) Up() (*bind.UniversalUDPMuxDefault, error) {
func (t *USPDevice) Up() (*udpmux.UniversalUDPMuxDefault, error) {
if t.device == nil {
return nil, fmt.Errorf("device is not ready yet")
}
@@ -118,6 +119,10 @@ func (t *USPDevice) WgAddress() wgaddr.Address {
return t.address
}
func (t *USPDevice) MTU() uint16 {
return t.mtu
}
func (t *USPDevice) DeviceName() string {
return t.name
}

View File

@@ -13,6 +13,7 @@ import (
"github.com/netbirdio/netbird/client/iface/bind"
"github.com/netbirdio/netbird/client/iface/configurer"
"github.com/netbirdio/netbird/client/iface/udpmux"
"github.com/netbirdio/netbird/client/iface/wgaddr"
)
@@ -23,17 +24,17 @@ type TunDevice struct {
address wgaddr.Address
port int
key string
mtu int
mtu uint16
iceBind *bind.ICEBind
device *device.Device
nativeTunDevice *tun.NativeTun
filteredDevice *FilteredDevice
udpMux *bind.UniversalUDPMuxDefault
udpMux *udpmux.UniversalUDPMuxDefault
configurer WGConfigurer
}
func NewTunDevice(name string, address wgaddr.Address, port int, key string, mtu int, iceBind *bind.ICEBind) *TunDevice {
func NewTunDevice(name string, address wgaddr.Address, port int, key string, mtu uint16, iceBind *bind.ICEBind) *TunDevice {
return &TunDevice{
name: name,
address: address,
@@ -59,7 +60,7 @@ func (t *TunDevice) Create() (WGConfigurer, error) {
return nil, err
}
log.Info("create tun interface")
tunDevice, err := tun.CreateTUNWithRequestedGUID(t.name, &guid, t.mtu)
tunDevice, err := tun.CreateTUNWithRequestedGUID(t.name, &guid, int(t.mtu))
if err != nil {
return nil, fmt.Errorf("error creating tun device: %s", err)
}
@@ -104,7 +105,7 @@ func (t *TunDevice) Create() (WGConfigurer, error) {
return t.configurer, nil
}
func (t *TunDevice) Up() (*bind.UniversalUDPMuxDefault, error) {
func (t *TunDevice) Up() (*udpmux.UniversalUDPMuxDefault, error) {
err := t.device.Up()
if err != nil {
return nil, err
@@ -144,6 +145,10 @@ func (t *TunDevice) WgAddress() wgaddr.Address {
return t.address
}
func (t *TunDevice) MTU() uint16 {
return t.mtu
}
func (t *TunDevice) DeviceName() string {
return t.name
}

View File

@@ -5,16 +5,17 @@ import (
"golang.zx2c4.com/wireguard/tun/netstack"
"github.com/netbirdio/netbird/client/iface/bind"
"github.com/netbirdio/netbird/client/iface/device"
"github.com/netbirdio/netbird/client/iface/udpmux"
"github.com/netbirdio/netbird/client/iface/wgaddr"
)
type WGTunDevice interface {
Create(routes []string, dns string, searchDomains []string) (device.WGConfigurer, error)
Up() (*bind.UniversalUDPMuxDefault, error)
Up() (*udpmux.UniversalUDPMuxDefault, error)
UpdateAddr(address wgaddr.Address) error
WgAddress() wgaddr.Address
MTU() uint16
DeviceName() string
Close() error
FilteredDevice() *device.FilteredDevice

View File

@@ -16,9 +16,9 @@ import (
wgdevice "golang.zx2c4.com/wireguard/device"
"github.com/netbirdio/netbird/client/errors"
"github.com/netbirdio/netbird/client/iface/bind"
"github.com/netbirdio/netbird/client/iface/configurer"
"github.com/netbirdio/netbird/client/iface/device"
"github.com/netbirdio/netbird/client/iface/udpmux"
"github.com/netbirdio/netbird/client/iface/wgaddr"
"github.com/netbirdio/netbird/client/iface/wgproxy"
"github.com/netbirdio/netbird/monotime"
@@ -26,6 +26,8 @@ import (
const (
DefaultMTU = 1280
MinMTU = 576
MaxMTU = 8192
DefaultWgPort = 51820
WgInterfaceDefault = configurer.WgInterfaceDefault
)
@@ -35,6 +37,17 @@ var (
ErrIfaceNotFound = fmt.Errorf("wireguard interface not found")
)
// ValidateMTU validates that MTU is within acceptable range
func ValidateMTU(mtu uint16) error {
if mtu < MinMTU {
return fmt.Errorf("MTU %d below minimum (%d bytes)", mtu, MinMTU)
}
if mtu > MaxMTU {
return fmt.Errorf("MTU %d exceeds maximum supported size (%d bytes)", mtu, MaxMTU)
}
return nil
}
type wgProxyFactory interface {
GetProxy() wgproxy.Proxy
Free() error
@@ -45,10 +58,10 @@ type WGIFaceOpts struct {
Address string
WGPort int
WGPrivKey string
MTU int
MTU uint16
MobileArgs *device.MobileIFaceArguments
TransportNet transport.Net
FilterFn bind.FilterFn
FilterFn udpmux.FilterFn
DisableDNS bool
}
@@ -82,6 +95,10 @@ func (w *WGIface) Address() wgaddr.Address {
return w.tun.WgAddress()
}
func (w *WGIface) MTU() uint16 {
return w.tun.MTU()
}
// ToInterface returns the net.Interface for the Wireguard interface
func (r *WGIface) ToInterface() *net.Interface {
name := r.tun.DeviceName()
@@ -97,7 +114,7 @@ func (r *WGIface) ToInterface() *net.Interface {
// Up configures a Wireguard interface
// The interface must exist before calling this method (e.g. call interface.Create() before)
func (w *WGIface) Up() (*bind.UniversalUDPMuxDefault, error) {
func (w *WGIface) Up() (*udpmux.UniversalUDPMuxDefault, error) {
w.mu.Lock()
defer w.mu.Unlock()

View File

@@ -3,6 +3,7 @@ package iface
import (
"github.com/netbirdio/netbird/client/iface/bind"
"github.com/netbirdio/netbird/client/iface/device"
"github.com/netbirdio/netbird/client/iface/netstack"
"github.com/netbirdio/netbird/client/iface/wgaddr"
"github.com/netbirdio/netbird/client/iface/wgproxy"
)
@@ -14,7 +15,16 @@ func NewWGIFace(opts WGIFaceOpts) (*WGIface, error) {
return nil, err
}
iceBind := bind.NewICEBind(opts.TransportNet, opts.FilterFn, wgAddress)
iceBind := bind.NewICEBind(opts.TransportNet, opts.FilterFn, wgAddress, opts.MTU)
if netstack.IsEnabled() {
wgIFace := &WGIface{
userspaceBind: true,
tun: device.NewNetstackDevice(opts.IFaceName, wgAddress, opts.WGPort, opts.WGPrivKey, opts.MTU, iceBind, netstack.ListenAddr()),
wgProxyFactory: wgproxy.NewUSPFactory(iceBind),
}
return wgIFace, nil
}
wgIFace := &WGIface{
userspaceBind: true,

View File

@@ -17,7 +17,7 @@ func NewWGIFace(opts WGIFaceOpts) (*WGIface, error) {
return nil, err
}
iceBind := bind.NewICEBind(opts.TransportNet, opts.FilterFn, wgAddress)
iceBind := bind.NewICEBind(opts.TransportNet, opts.FilterFn, wgAddress, opts.MTU)
var tun WGTunDevice
if netstack.IsEnabled() {

View File

@@ -16,10 +16,10 @@ func NewWGIFace(opts WGIFaceOpts) (*WGIface, error) {
return nil, err
}
iceBind := bind.NewICEBind(opts.TransportNet, opts.FilterFn, wgAddress)
iceBind := bind.NewICEBind(opts.TransportNet, opts.FilterFn, wgAddress, opts.MTU)
wgIFace := &WGIface{
tun: device.NewTunDevice(opts.IFaceName, wgAddress, opts.WGPort, opts.WGPrivKey, iceBind, opts.MobileArgs.TunFd),
tun: device.NewTunDevice(opts.IFaceName, wgAddress, opts.WGPort, opts.WGPrivKey, opts.MTU, iceBind, opts.MobileArgs.TunFd),
userspaceBind: true,
wgProxyFactory: wgproxy.NewUSPFactory(iceBind),
}

View File

@@ -22,7 +22,7 @@ func NewWGIFace(opts WGIFaceOpts) (*WGIface, error) {
wgIFace := &WGIface{}
if netstack.IsEnabled() {
iceBind := bind.NewICEBind(opts.TransportNet, opts.FilterFn, wgAddress)
iceBind := bind.NewICEBind(opts.TransportNet, opts.FilterFn, wgAddress, opts.MTU)
wgIFace.tun = device.NewNetstackDevice(opts.IFaceName, wgAddress, opts.WGPort, opts.WGPrivKey, opts.MTU, iceBind, netstack.ListenAddr())
wgIFace.userspaceBind = true
wgIFace.wgProxyFactory = wgproxy.NewUSPFactory(iceBind)
@@ -31,11 +31,11 @@ func NewWGIFace(opts WGIFaceOpts) (*WGIface, error) {
if device.WireGuardModuleIsLoaded() {
wgIFace.tun = device.NewKernelDevice(opts.IFaceName, wgAddress, opts.WGPort, opts.WGPrivKey, opts.MTU, opts.TransportNet)
wgIFace.wgProxyFactory = wgproxy.NewKernelFactory(opts.WGPort)
wgIFace.wgProxyFactory = wgproxy.NewKernelFactory(opts.WGPort, opts.MTU)
return wgIFace, nil
}
if device.ModuleTunIsLoaded() {
iceBind := bind.NewICEBind(opts.TransportNet, opts.FilterFn, wgAddress)
iceBind := bind.NewICEBind(opts.TransportNet, opts.FilterFn, wgAddress, opts.MTU)
wgIFace.tun = device.NewUSPDevice(opts.IFaceName, wgAddress, opts.WGPort, opts.WGPrivKey, opts.MTU, iceBind)
wgIFace.userspaceBind = true
wgIFace.wgProxyFactory = wgproxy.NewUSPFactory(iceBind)

View File

@@ -14,7 +14,7 @@ func NewWGIFace(opts WGIFaceOpts) (*WGIface, error) {
if err != nil {
return nil, err
}
iceBind := bind.NewICEBind(opts.TransportNet, opts.FilterFn, wgAddress)
iceBind := bind.NewICEBind(opts.TransportNet, opts.FilterFn, wgAddress, opts.MTU)
var tun WGTunDevice
if netstack.IsEnabled() {

View File

@@ -1,4 +1,4 @@
package bind
package udpmux
/*
Most of this code was copied from https://github.com/pion/ice and modified to fulfill NetBird's requirements
@@ -16,11 +16,12 @@ import (
)
type udpMuxedConnParams struct {
Mux *UDPMuxDefault
AddrPool *sync.Pool
Key string
LocalAddr net.Addr
Logger logging.LeveledLogger
Mux *SingleSocketUDPMux
AddrPool *sync.Pool
Key string
LocalAddr net.Addr
Logger logging.LeveledLogger
CandidateID string
}
// udpMuxedConn represents a logical packet conn for a single remote as identified by ufrag
@@ -119,6 +120,10 @@ func (c *udpMuxedConn) Close() error {
return err
}
func (c *udpMuxedConn) GetCandidateID() string {
return c.params.CandidateID
}
func (c *udpMuxedConn) isClosed() bool {
select {
case <-c.closedChan:

View File

@@ -0,0 +1,64 @@
// Package udpmux provides a custom implementation of a UDP multiplexer
// that allows multiple logical ICE connections to share a single underlying
// UDP socket. This is based on Pion's ICE library, with modifications for
// NetBird's requirements.
//
// # Background
//
// In WebRTC and NAT traversal scenarios, ICE (Interactive Connectivity
// Establishment) is responsible for discovering candidate network paths
// and maintaining connectivity between peers. Each ICE connection
// normally requires a dedicated UDP socket. However, using one socket
// per candidate can be inefficient and difficult to manage.
//
// This package introduces SingleSocketUDPMux, which allows multiple ICE
// candidate connections (muxed connections) to share a single UDP socket.
// It handles demultiplexing of packets based on ICE ufrag values, STUN
// attributes, and candidate IDs.
//
// # Usage
//
// The typical flow is:
//
// 1. Create a UDP socket (net.PacketConn).
// 2. Construct Params with the socket and optional logger/net stack.
// 3. Call NewSingleSocketUDPMux(params).
// 4. For each ICE candidate ufrag, call GetConn(ufrag, addr, candidateID)
// to obtain a logical PacketConn.
// 5. Use the returned PacketConn just like a normal UDP connection.
//
// # STUN Message Routing Logic
//
// When a STUN packet arrives, the mux decides which connection should
// receive it using this routing logic:
//
// Primary Routing: Candidate Pair ID
// - Extract the candidate pair ID from the STUN message using
// ice.CandidatePairIDFromSTUN(msg)
// - The target candidate is the locally generated candidate that
// corresponds to the connection that should handle this STUN message
// - If found, use the target candidate ID to lookup the specific
// connection in candidateConnMap
// - Route the message directly to that connection
//
// Fallback Routing: Broadcasting
// When candidate pair ID is not available or lookup fails:
// - Collect connections from addressMap based on source address
// - Find connection using username attribute (ufrag) from STUN message
// - Remove duplicate connections from the list
// - Send the STUN message to all collected connections
//
// # Peer Reflexive Candidate Discovery
//
// When a remote peer sends a STUN message from an unknown source address
// (from a candidate that has not been exchanged via signal), the ICE
// library will:
// - Generate a new peer reflexive candidate for this source address
// - Extract or assign a candidate ID based on the STUN message attributes
// - Create a mapping between the new peer reflexive candidate ID and
// the appropriate local connection
//
// This discovery mechanism ensures that STUN messages from newly discovered
// peer reflexive candidates can be properly routed to the correct local
// connection without requiring fallback broadcasting.
package udpmux

View File

@@ -1,4 +1,4 @@
package bind
package udpmux
import (
"fmt"
@@ -8,9 +8,9 @@ import (
"strings"
"sync"
"github.com/pion/ice/v3"
"github.com/pion/ice/v4"
"github.com/pion/logging"
"github.com/pion/stun/v2"
"github.com/pion/stun/v3"
"github.com/pion/transport/v3"
"github.com/pion/transport/v3/stdnet"
log "github.com/sirupsen/logrus"
@@ -22,9 +22,9 @@ import (
const receiveMTU = 8192
// UDPMuxDefault is an implementation of the interface
type UDPMuxDefault struct {
params UDPMuxParams
// SingleSocketUDPMux is an implementation of the interface
type SingleSocketUDPMux struct {
params Params
closedChan chan struct{}
closeOnce sync.Once
@@ -32,6 +32,9 @@ type UDPMuxDefault struct {
// connsIPv4 and connsIPv6 are maps of all udpMuxedConn indexed by ufrag|network|candidateType
connsIPv4, connsIPv6 map[string]*udpMuxedConn
// candidateConnMap maps local candidate IDs to their corresponding connection.
candidateConnMap map[string]*udpMuxedConn
addressMapMu sync.RWMutex
addressMap map[string][]*udpMuxedConn
@@ -46,8 +49,8 @@ type UDPMuxDefault struct {
const maxAddrSize = 512
// UDPMuxParams are parameters for UDPMux.
type UDPMuxParams struct {
// Params are parameters for UDPMux.
type Params struct {
Logger logging.LeveledLogger
UDPConn net.PacketConn
@@ -147,18 +150,19 @@ func isZeros(ip net.IP) bool {
return true
}
// NewUDPMuxDefault creates an implementation of UDPMux
func NewUDPMuxDefault(params UDPMuxParams) *UDPMuxDefault {
// NewSingleSocketUDPMux creates an implementation of UDPMux
func NewSingleSocketUDPMux(params Params) *SingleSocketUDPMux {
if params.Logger == nil {
params.Logger = getLogger()
}
mux := &UDPMuxDefault{
addressMap: map[string][]*udpMuxedConn{},
params: params,
connsIPv4: make(map[string]*udpMuxedConn),
connsIPv6: make(map[string]*udpMuxedConn),
closedChan: make(chan struct{}, 1),
mux := &SingleSocketUDPMux{
addressMap: map[string][]*udpMuxedConn{},
params: params,
connsIPv4: make(map[string]*udpMuxedConn),
connsIPv6: make(map[string]*udpMuxedConn),
candidateConnMap: make(map[string]*udpMuxedConn),
closedChan: make(chan struct{}, 1),
pool: &sync.Pool{
New: func() interface{} {
// big enough buffer to fit both packet and address
@@ -171,15 +175,15 @@ func NewUDPMuxDefault(params UDPMuxParams) *UDPMuxDefault {
return mux
}
func (m *UDPMuxDefault) updateLocalAddresses() {
func (m *SingleSocketUDPMux) updateLocalAddresses() {
var localAddrsForUnspecified []net.Addr
if addr, ok := m.params.UDPConn.LocalAddr().(*net.UDPAddr); !ok {
m.params.Logger.Errorf("LocalAddr is not a net.UDPAddr, got %T", m.params.UDPConn.LocalAddr())
} else if ok && addr.IP.IsUnspecified() {
// For unspecified addresses, the correct behavior is to return errListenUnspecified, but
// it will break the applications that are already using unspecified UDP connection
// with UDPMuxDefault, so print a warn log and create a local address list for mux.
m.params.Logger.Warn("UDPMuxDefault should not listening on unspecified address, use NewMultiUDPMuxFromPort instead")
// with SingleSocketUDPMux, so print a warn log and create a local address list for mux.
m.params.Logger.Warn("SingleSocketUDPMux should not listening on unspecified address, use NewMultiUDPMuxFromPort instead")
var networks []ice.NetworkType
switch {
@@ -216,13 +220,13 @@ func (m *UDPMuxDefault) updateLocalAddresses() {
m.mu.Unlock()
}
// LocalAddr returns the listening address of this UDPMuxDefault
func (m *UDPMuxDefault) LocalAddr() net.Addr {
// LocalAddr returns the listening address of this SingleSocketUDPMux
func (m *SingleSocketUDPMux) LocalAddr() net.Addr {
return m.params.UDPConn.LocalAddr()
}
// GetListenAddresses returns the list of addresses that this mux is listening on
func (m *UDPMuxDefault) GetListenAddresses() []net.Addr {
func (m *SingleSocketUDPMux) GetListenAddresses() []net.Addr {
m.updateLocalAddresses()
m.mu.Lock()
@@ -236,7 +240,7 @@ func (m *UDPMuxDefault) GetListenAddresses() []net.Addr {
// GetConn returns a PacketConn given the connection's ufrag and network address
// creates the connection if an existing one can't be found
func (m *UDPMuxDefault) GetConn(ufrag string, addr net.Addr) (net.PacketConn, error) {
func (m *SingleSocketUDPMux) GetConn(ufrag string, addr net.Addr, candidateID string) (net.PacketConn, error) {
// don't check addr for mux using unspecified address
m.mu.Lock()
lenLocalAddrs := len(m.localAddrsForUnspecified)
@@ -260,12 +264,14 @@ func (m *UDPMuxDefault) GetConn(ufrag string, addr net.Addr) (net.PacketConn, er
return conn, nil
}
c := m.createMuxedConn(ufrag)
c := m.createMuxedConn(ufrag, candidateID)
go func() {
<-c.CloseChannel()
m.RemoveConnByUfrag(ufrag)
}()
m.candidateConnMap[candidateID] = c
if isIPv6 {
m.connsIPv6[ufrag] = c
} else {
@@ -276,7 +282,7 @@ func (m *UDPMuxDefault) GetConn(ufrag string, addr net.Addr) (net.PacketConn, er
}
// RemoveConnByUfrag stops and removes the muxed packet connection
func (m *UDPMuxDefault) RemoveConnByUfrag(ufrag string) {
func (m *SingleSocketUDPMux) RemoveConnByUfrag(ufrag string) {
removedConns := make([]*udpMuxedConn, 0, 2)
// Keep lock section small to avoid deadlock with conn lock
@@ -284,10 +290,12 @@ func (m *UDPMuxDefault) RemoveConnByUfrag(ufrag string) {
if c, ok := m.connsIPv4[ufrag]; ok {
delete(m.connsIPv4, ufrag)
removedConns = append(removedConns, c)
delete(m.candidateConnMap, c.GetCandidateID())
}
if c, ok := m.connsIPv6[ufrag]; ok {
delete(m.connsIPv6, ufrag)
removedConns = append(removedConns, c)
delete(m.candidateConnMap, c.GetCandidateID())
}
m.mu.Unlock()
@@ -314,7 +322,7 @@ func (m *UDPMuxDefault) RemoveConnByUfrag(ufrag string) {
}
// IsClosed returns true if the mux had been closed
func (m *UDPMuxDefault) IsClosed() bool {
func (m *SingleSocketUDPMux) IsClosed() bool {
select {
case <-m.closedChan:
return true
@@ -324,7 +332,7 @@ func (m *UDPMuxDefault) IsClosed() bool {
}
// Close the mux, no further connections could be created
func (m *UDPMuxDefault) Close() error {
func (m *SingleSocketUDPMux) Close() error {
var err error
m.closeOnce.Do(func() {
m.mu.Lock()
@@ -347,11 +355,11 @@ func (m *UDPMuxDefault) Close() error {
return err
}
func (m *UDPMuxDefault) writeTo(buf []byte, rAddr net.Addr) (n int, err error) {
func (m *SingleSocketUDPMux) writeTo(buf []byte, rAddr net.Addr) (n int, err error) {
return m.params.UDPConn.WriteTo(buf, rAddr)
}
func (m *UDPMuxDefault) registerConnForAddress(conn *udpMuxedConn, addr string) {
func (m *SingleSocketUDPMux) registerConnForAddress(conn *udpMuxedConn, addr string) {
if m.IsClosed() {
return
}
@@ -368,81 +376,109 @@ func (m *UDPMuxDefault) registerConnForAddress(conn *udpMuxedConn, addr string)
log.Debugf("ICE: registered %s for %s", addr, conn.params.Key)
}
func (m *UDPMuxDefault) createMuxedConn(key string) *udpMuxedConn {
func (m *SingleSocketUDPMux) createMuxedConn(key string, candidateID string) *udpMuxedConn {
c := newUDPMuxedConn(&udpMuxedConnParams{
Mux: m,
Key: key,
AddrPool: m.pool,
LocalAddr: m.LocalAddr(),
Logger: m.params.Logger,
Mux: m,
Key: key,
AddrPool: m.pool,
LocalAddr: m.LocalAddr(),
Logger: m.params.Logger,
CandidateID: candidateID,
})
return c
}
// HandleSTUNMessage handles STUN packets and forwards them to underlying pion/ice library
func (m *UDPMuxDefault) HandleSTUNMessage(msg *stun.Message, addr net.Addr) error {
func (m *SingleSocketUDPMux) HandleSTUNMessage(msg *stun.Message, addr net.Addr) error {
remoteAddr, ok := addr.(*net.UDPAddr)
if !ok {
return fmt.Errorf("underlying PacketConn did not return a UDPAddr")
}
// If we have already seen this address dispatch to the appropriate destination
// If you are using the same socket for the Host and SRFLX candidates, it might be that there are more than one
// muxed connection - one for the SRFLX candidate and the other one for the HOST one.
// We will then forward STUN packets to each of these connections.
m.addressMapMu.RLock()
// Try to route to specific candidate connection first
if conn := m.findCandidateConnection(msg); conn != nil {
return conn.writePacket(msg.Raw, remoteAddr)
}
// Fallback: route to all possible connections
return m.forwardToAllConnections(msg, addr, remoteAddr)
}
// findCandidateConnection attempts to find the specific connection for a STUN message
func (m *SingleSocketUDPMux) findCandidateConnection(msg *stun.Message) *udpMuxedConn {
candidatePairID, ok, err := ice.CandidatePairIDFromSTUN(msg)
if err != nil {
return nil
} else if !ok {
return nil
}
m.mu.Lock()
defer m.mu.Unlock()
conn, exists := m.candidateConnMap[candidatePairID.TargetCandidateID()]
if !exists {
return nil
}
return conn
}
// forwardToAllConnections forwards STUN message to all relevant connections
func (m *SingleSocketUDPMux) forwardToAllConnections(msg *stun.Message, addr net.Addr, remoteAddr *net.UDPAddr) error {
var destinationConnList []*udpMuxedConn
// Add connections from address map
m.addressMapMu.RLock()
if storedConns, ok := m.addressMap[addr.String()]; ok {
destinationConnList = append(destinationConnList, storedConns...)
}
m.addressMapMu.RUnlock()
var isIPv6 bool
if udpAddr, _ := addr.(*net.UDPAddr); udpAddr != nil && udpAddr.IP.To4() == nil {
isIPv6 = true
if conn, ok := m.findConnectionByUsername(msg, addr); ok {
// If we have already seen this address dispatch to the appropriate destination
// If you are using the same socket for the Host and SRFLX candidates, it might be that there are more than one
// muxed connection - one for the SRFLX candidate and the other one for the HOST one.
// We will then forward STUN packets to each of these connections.
if !m.connectionExists(conn, destinationConnList) {
destinationConnList = append(destinationConnList, conn)
}
}
// This block is needed to discover Peer Reflexive Candidates for which we don't know the Endpoint upfront.
// However, we can take a username attribute from the STUN message which contains ufrag.
// We can use ufrag to identify the destination conn to route packet to.
attr, stunAttrErr := msg.Get(stun.AttrUsername)
if stunAttrErr == nil {
ufrag := strings.Split(string(attr), ":")[0]
m.mu.Lock()
destinationConn := m.connsIPv4[ufrag]
if isIPv6 {
destinationConn = m.connsIPv6[ufrag]
}
if destinationConn != nil {
exists := false
for _, conn := range destinationConnList {
if conn.params.Key == destinationConn.params.Key {
exists = true
break
}
}
if !exists {
destinationConnList = append(destinationConnList, destinationConn)
}
}
m.mu.Unlock()
}
// Forward STUN packets to each destination connections even thought the STUN packet might not belong there.
// It will be discarded by the further ICE candidate logic if so.
// Forward to all found connections
for _, conn := range destinationConnList {
if err := conn.writePacket(msg.Raw, remoteAddr); err != nil {
log.Errorf("could not write packet: %v", err)
}
}
return nil
}
func (m *UDPMuxDefault) getConn(ufrag string, isIPv6 bool) (val *udpMuxedConn, ok bool) {
// findConnectionByUsername finds connection using username attribute from STUN message
func (m *SingleSocketUDPMux) findConnectionByUsername(msg *stun.Message, addr net.Addr) (*udpMuxedConn, bool) {
attr, err := msg.Get(stun.AttrUsername)
if err != nil {
return nil, false
}
ufrag := strings.Split(string(attr), ":")[0]
isIPv6 := isIPv6Address(addr)
m.mu.Lock()
defer m.mu.Unlock()
return m.getConn(ufrag, isIPv6)
}
// connectionExists checks if a connection already exists in the list
func (m *SingleSocketUDPMux) connectionExists(target *udpMuxedConn, conns []*udpMuxedConn) bool {
for _, conn := range conns {
if conn.params.Key == target.params.Key {
return true
}
}
return false
}
func (m *SingleSocketUDPMux) getConn(ufrag string, isIPv6 bool) (val *udpMuxedConn, ok bool) {
if isIPv6 {
val, ok = m.connsIPv6[ufrag]
} else {
@@ -451,6 +487,13 @@ func (m *UDPMuxDefault) getConn(ufrag string, isIPv6 bool) (val *udpMuxedConn, o
return
}
func isIPv6Address(addr net.Addr) bool {
if udpAddr, ok := addr.(*net.UDPAddr); ok {
return udpAddr.IP.To4() == nil
}
return false
}
type bufferHolder struct {
buf []byte
}

View File

@@ -1,12 +1,12 @@
//go:build !ios
package bind
package udpmux
import (
nbnet "github.com/netbirdio/netbird/util/net"
nbnet "github.com/netbirdio/netbird/client/net"
)
func (m *UDPMuxDefault) notifyAddressRemoval(addr string) {
func (m *SingleSocketUDPMux) notifyAddressRemoval(addr string) {
// Kernel mode: direct nbnet.PacketConn (SharedSocket wrapped with nbnet)
if conn, ok := m.params.UDPConn.(*nbnet.PacketConn); ok {
conn.RemoveAddress(addr)

View File

@@ -0,0 +1,7 @@
//go:build ios
package udpmux
func (m *SingleSocketUDPMux) notifyAddressRemoval(addr string) {
// iOS doesn't support nbnet hooks, so this is a no-op
}

View File

@@ -1,4 +1,4 @@
package bind
package udpmux
/*
Most of this code was copied from https://github.com/pion/ice and modified to fulfill NetBird's requirements.
@@ -15,9 +15,10 @@ import (
log "github.com/sirupsen/logrus"
"github.com/pion/logging"
"github.com/pion/stun/v2"
"github.com/pion/stun/v3"
"github.com/pion/transport/v3"
"github.com/netbirdio/netbird/client/iface/bufsize"
"github.com/netbirdio/netbird/client/iface/wgaddr"
)
@@ -28,7 +29,7 @@ type FilterFn func(address netip.Addr) (bool, netip.Prefix, error)
// UniversalUDPMuxDefault handles STUN and TURN servers packets by wrapping the original UDPConn
// It then passes packets to the UDPMux that does the actual connection muxing.
type UniversalUDPMuxDefault struct {
*UDPMuxDefault
*SingleSocketUDPMux
params UniversalUDPMuxParams
// since we have a shared socket, for srflx candidates it makes sense to have a shared mapped address across all the agents
@@ -44,6 +45,7 @@ type UniversalUDPMuxParams struct {
Net transport.Net
FilterFn FilterFn
WGAddress wgaddr.Address
MTU uint16
}
// NewUniversalUDPMuxDefault creates an implementation of UniversalUDPMux embedding UDPMux
@@ -70,12 +72,12 @@ func NewUniversalUDPMuxDefault(params UniversalUDPMuxParams) *UniversalUDPMuxDef
address: params.WGAddress,
}
udpMuxParams := UDPMuxParams{
udpMuxParams := Params{
Logger: params.Logger,
UDPConn: m.params.UDPConn,
Net: m.params.Net,
}
m.UDPMuxDefault = NewUDPMuxDefault(udpMuxParams)
m.SingleSocketUDPMux = NewSingleSocketUDPMux(udpMuxParams)
return m
}
@@ -84,7 +86,7 @@ func NewUniversalUDPMuxDefault(params UniversalUDPMuxParams) *UniversalUDPMuxDef
// just ignore other packets printing an warning message.
// It is a blocking method, consider running in a go routine.
func (m *UniversalUDPMuxDefault) ReadFromConn(ctx context.Context) {
buf := make([]byte, 1500)
buf := make([]byte, m.params.MTU+bufsize.WGBufferOverhead)
for {
select {
case <-ctx.Done():
@@ -209,8 +211,8 @@ func (m *UniversalUDPMuxDefault) GetRelayedAddr(turnAddr net.Addr, deadline time
// GetConnForURL add uniques to the muxed connection by concatenating ufrag and URL (e.g. STUN URL) to be able to support multiple STUN/TURN servers
// and return a unique connection per server.
func (m *UniversalUDPMuxDefault) GetConnForURL(ufrag string, url string, addr net.Addr) (net.PacketConn, error) {
return m.UDPMuxDefault.GetConn(fmt.Sprintf("%s%s", ufrag, url), addr)
func (m *UniversalUDPMuxDefault) GetConnForURL(ufrag string, url string, addr net.Addr, candidateID string) (net.PacketConn, error) {
return m.SingleSocketUDPMux.GetConn(fmt.Sprintf("%s%s", ufrag, url), addr, candidateID)
}
// HandleSTUNMessage discovers STUN packets that carry a XOR mapped address from a STUN server.
@@ -231,7 +233,7 @@ func (m *UniversalUDPMuxDefault) HandleSTUNMessage(msg *stun.Message, addr net.A
}
return nil
}
return m.UDPMuxDefault.HandleSTUNMessage(msg, addr)
return m.SingleSocketUDPMux.HandleSTUNMessage(msg, addr)
}
// isXORMappedResponse indicates whether the message is a XORMappedAddress and is coming from the known STUN server.

View File

@@ -12,6 +12,7 @@ import (
log "github.com/sirupsen/logrus"
"github.com/netbirdio/netbird/client/iface/bind"
"github.com/netbirdio/netbird/client/iface/bufsize"
"github.com/netbirdio/netbird/client/iface/wgproxy/listener"
)
@@ -135,7 +136,7 @@ func (p *ProxyBind) proxyToLocal(ctx context.Context) {
}()
for {
buf := make([]byte, 1500)
buf := make([]byte, p.Bind.MTU()+bufsize.WGBufferOverhead)
n, err := p.remoteConn.Read(buf)
if err != nil {
if ctx.Err() != nil {

View File

@@ -17,9 +17,10 @@ import (
log "github.com/sirupsen/logrus"
nberrors "github.com/netbirdio/netbird/client/errors"
"github.com/netbirdio/netbird/client/iface/bufsize"
"github.com/netbirdio/netbird/client/internal/ebpf"
ebpfMgr "github.com/netbirdio/netbird/client/internal/ebpf/manager"
nbnet "github.com/netbirdio/netbird/util/net"
nbnet "github.com/netbirdio/netbird/client/net"
)
const (
@@ -29,6 +30,7 @@ const (
// WGEBPFProxy definition for proxy with EBPF support
type WGEBPFProxy struct {
localWGListenPort int
mtu uint16
ebpfManager ebpfMgr.Manager
turnConnStore map[uint16]net.Conn
@@ -43,10 +45,11 @@ type WGEBPFProxy struct {
}
// NewWGEBPFProxy create new WGEBPFProxy instance
func NewWGEBPFProxy(wgPort int) *WGEBPFProxy {
func NewWGEBPFProxy(wgPort int, mtu uint16) *WGEBPFProxy {
log.Debugf("instantiate ebpf proxy")
wgProxy := &WGEBPFProxy{
localWGListenPort: wgPort,
mtu: mtu,
ebpfManager: ebpf.GetEbpfManagerInstance(),
turnConnStore: make(map[uint16]net.Conn),
}
@@ -138,7 +141,7 @@ func (p *WGEBPFProxy) Free() error {
// proxyToRemote read messages from local WireGuard interface and forward it to remote conn
// From this go routine has only one instance.
func (p *WGEBPFProxy) proxyToRemote() {
buf := make([]byte, 1500)
buf := make([]byte, p.mtu+bufsize.WGBufferOverhead)
for p.ctx.Err() == nil {
if err := p.readAndForwardPacket(buf); err != nil {
if p.ctx.Err() != nil {

View File

@@ -7,7 +7,7 @@ import (
)
func TestWGEBPFProxy_connStore(t *testing.T) {
wgProxy := NewWGEBPFProxy(1)
wgProxy := NewWGEBPFProxy(1, 1280)
p, _ := wgProxy.storeTurnConn(nil)
if p != 1 {
@@ -27,7 +27,7 @@ func TestWGEBPFProxy_connStore(t *testing.T) {
}
func TestWGEBPFProxy_portCalculation_overflow(t *testing.T) {
wgProxy := NewWGEBPFProxy(1)
wgProxy := NewWGEBPFProxy(1, 1280)
_, _ = wgProxy.storeTurnConn(nil)
wgProxy.lastUsedPort = 65535
@@ -43,7 +43,7 @@ func TestWGEBPFProxy_portCalculation_overflow(t *testing.T) {
}
func TestWGEBPFProxy_portCalculation_maxConn(t *testing.T) {
wgProxy := NewWGEBPFProxy(1)
wgProxy := NewWGEBPFProxy(1, 1280)
for i := 0; i < 65535; i++ {
_, _ = wgProxy.storeTurnConn(nil)

View File

@@ -12,6 +12,7 @@ import (
log "github.com/sirupsen/logrus"
"github.com/netbirdio/netbird/client/iface/bufsize"
"github.com/netbirdio/netbird/client/iface/wgproxy/listener"
)
@@ -103,7 +104,7 @@ func (e *ProxyWrapper) CloseConn() error {
func (p *ProxyWrapper) proxyToLocal(ctx context.Context) {
defer p.WgeBPFProxy.removeTurnConn(uint16(p.wgEndpointAddr.Port))
buf := make([]byte, 1500)
buf := make([]byte, p.WgeBPFProxy.mtu+bufsize.WGBufferOverhead)
for {
n, err := p.readFromRemote(ctx, buf)
if err != nil {

View File

@@ -11,16 +11,18 @@ import (
type KernelFactory struct {
wgPort int
mtu uint16
ebpfProxy *ebpf.WGEBPFProxy
}
func NewKernelFactory(wgPort int) *KernelFactory {
func NewKernelFactory(wgPort int, mtu uint16) *KernelFactory {
f := &KernelFactory{
wgPort: wgPort,
mtu: mtu,
}
ebpfProxy := ebpf.NewWGEBPFProxy(wgPort)
ebpfProxy := ebpf.NewWGEBPFProxy(wgPort, mtu)
if err := ebpfProxy.Listen(); err != nil {
log.Infof("WireGuard Proxy Factory will produce UDP proxy")
log.Warnf("failed to initialize ebpf proxy, fallback to user space proxy: %s", err)
@@ -33,7 +35,7 @@ func NewKernelFactory(wgPort int) *KernelFactory {
func (w *KernelFactory) GetProxy() Proxy {
if w.ebpfProxy == nil {
return udpProxy.NewWGUDPProxy(w.wgPort)
return udpProxy.NewWGUDPProxy(w.wgPort, w.mtu)
}
return ebpf.NewProxyWrapper(w.ebpfProxy)

View File

@@ -9,19 +9,21 @@ import (
// KernelFactory todo: check eBPF support on FreeBSD
type KernelFactory struct {
wgPort int
mtu uint16
}
func NewKernelFactory(wgPort int) *KernelFactory {
func NewKernelFactory(wgPort int, mtu uint16) *KernelFactory {
log.Infof("WireGuard Proxy Factory will produce UDP proxy")
f := &KernelFactory{
wgPort: wgPort,
mtu: mtu,
}
return f
}
func (w *KernelFactory) GetProxy() Proxy {
return udpProxy.NewWGUDPProxy(w.wgPort)
return udpProxy.NewWGUDPProxy(w.wgPort, w.mtu)
}
func (w *KernelFactory) Free() error {

View File

@@ -16,7 +16,7 @@ func TestProxyCloseByRemoteConnEBPF(t *testing.T) {
}
ctx := context.Background()
ebpfProxy := ebpf.NewWGEBPFProxy(51831)
ebpfProxy := ebpf.NewWGEBPFProxy(51831, 1280)
if err := ebpfProxy.Listen(); err != nil {
t.Fatalf("failed to initialize ebpf proxy: %s", err)
}

View File

@@ -84,12 +84,12 @@ func TestProxyCloseByRemoteConn(t *testing.T) {
}{
{
name: "userspace proxy",
proxy: udpProxy.NewWGUDPProxy(51830),
proxy: udpProxy.NewWGUDPProxy(51830, 1280),
},
}
if runtime.GOOS == "linux" && os.Getenv("GITHUB_ACTIONS") != "true" {
ebpfProxy := ebpf.NewWGEBPFProxy(51831)
ebpfProxy := ebpf.NewWGEBPFProxy(51831, 1280)
if err := ebpfProxy.Listen(); err != nil {
t.Fatalf("failed to initialize ebpf proxy: %s", err)
}

View File

@@ -12,12 +12,14 @@ import (
log "github.com/sirupsen/logrus"
cerrors "github.com/netbirdio/netbird/client/errors"
"github.com/netbirdio/netbird/client/iface/bufsize"
"github.com/netbirdio/netbird/client/iface/wgproxy/listener"
)
// WGUDPProxy proxies
type WGUDPProxy struct {
localWGListenPort int
mtu uint16
remoteConn net.Conn
localConn net.Conn
@@ -34,10 +36,11 @@ type WGUDPProxy struct {
}
// NewWGUDPProxy instantiate a UDP based WireGuard proxy. This is not a thread safe implementation
func NewWGUDPProxy(wgPort int) *WGUDPProxy {
func NewWGUDPProxy(wgPort int, mtu uint16) *WGUDPProxy {
log.Debugf("Initializing new user space proxy with port %d", wgPort)
p := &WGUDPProxy{
localWGListenPort: wgPort,
mtu: mtu,
closeListener: listener.NewCloseListener(),
}
return p
@@ -144,7 +147,7 @@ func (p *WGUDPProxy) proxyToRemote(ctx context.Context) {
}
}()
buf := make([]byte, 1500)
buf := make([]byte, p.mtu+bufsize.WGBufferOverhead)
for ctx.Err() == nil {
n, err := p.localConn.Read(buf)
if err != nil {
@@ -179,7 +182,7 @@ func (p *WGUDPProxy) proxyToLocal(ctx context.Context) {
}
}()
buf := make([]byte, 1500)
buf := make([]byte, p.mtu+bufsize.WGBufferOverhead)
for {
n, err := p.remoteConnRead(ctx, buf)
if err != nil {

View File

@@ -3,15 +3,17 @@ package auth
import (
"context"
"fmt"
"github.com/golang-jwt/jwt"
"github.com/netbirdio/netbird/client/internal"
"github.com/stretchr/testify/require"
"io"
"net/http"
"net/url"
"strings"
"testing"
"time"
"github.com/golang-jwt/jwt/v5"
"github.com/stretchr/testify/require"
"github.com/netbirdio/netbird/client/internal"
)
type mockHTTPClient struct {

View File

@@ -18,6 +18,7 @@ import (
"google.golang.org/grpc/codes"
gstatus "google.golang.org/grpc/status"
"github.com/netbirdio/netbird/client/iface"
"github.com/netbirdio/netbird/client/iface/device"
"github.com/netbirdio/netbird/client/internal/dns"
"github.com/netbirdio/netbird/client/internal/listener"
@@ -33,7 +34,7 @@ import (
relayClient "github.com/netbirdio/netbird/shared/relay/client"
signal "github.com/netbirdio/netbird/shared/signal/client"
"github.com/netbirdio/netbird/util"
nbnet "github.com/netbirdio/netbird/util/net"
nbnet "github.com/netbirdio/netbird/client/net"
"github.com/netbirdio/netbird/version"
)
@@ -244,7 +245,15 @@ func (c *ConnectClient) run(mobileDependency MobileDependency, runningChan chan
c.statusRecorder.MarkSignalConnected()
relayURLs, token := parseRelayInfo(loginResp)
relayManager := relayClient.NewManager(engineCtx, relayURLs, myPrivateKey.PublicKey().String())
peerConfig := loginResp.GetPeerConfig()
engineConfig, err := createEngineConfig(myPrivateKey, c.config, peerConfig)
if err != nil {
log.Error(err)
return wrapErr(err)
}
relayManager := relayClient.NewManager(engineCtx, relayURLs, myPrivateKey.PublicKey().String(), engineConfig.MTU)
c.statusRecorder.SetRelayMgr(relayManager)
if len(relayURLs) > 0 {
if token != nil {
@@ -259,14 +268,6 @@ func (c *ConnectClient) run(mobileDependency MobileDependency, runningChan chan
}
}
peerConfig := loginResp.GetPeerConfig()
engineConfig, err := createEngineConfig(myPrivateKey, c.config, peerConfig)
if err != nil {
log.Error(err)
return wrapErr(err)
}
checks := loginResp.GetChecks()
c.engineMutex.Lock()
@@ -274,7 +275,7 @@ func (c *ConnectClient) run(mobileDependency MobileDependency, runningChan chan
c.engine.SetSyncResponsePersistence(c.persistSyncResponse)
c.engineMutex.Unlock()
if err := c.engine.Start(); err != nil {
if err := c.engine.Start(loginResp.GetNetbirdConfig(), c.config.ManagementURL); err != nil {
log.Errorf("error while starting Netbird Connection Engine: %s", err)
return wrapErr(err)
}
@@ -283,10 +284,8 @@ func (c *ConnectClient) run(mobileDependency MobileDependency, runningChan chan
state.Set(StatusConnected)
if runningChan != nil {
select {
case runningChan <- struct{}{}:
default:
}
close(runningChan)
runningChan = nil
}
<-engineCtx.Done()
@@ -444,6 +443,8 @@ func createEngineConfig(key wgtypes.Key, config *profilemanager.Config, peerConf
BlockInbound: config.BlockInbound,
LazyConnectionEnabled: config.LazyConnectionEnabled,
MTU: selectMTU(config.MTU, peerConfig.Mtu),
}
if config.PreSharedKey != "" {
@@ -466,6 +467,20 @@ func createEngineConfig(key wgtypes.Key, config *profilemanager.Config, peerConf
return engineConf, nil
}
func selectMTU(localMTU uint16, peerMTU int32) uint16 {
var finalMTU uint16 = iface.DefaultMTU
if localMTU > 0 {
finalMTU = localMTU
} else if peerMTU > 0 {
finalMTU = uint16(peerMTU)
}
// Set global DNS MTU
dns.SetCurrentMTU(finalMTU)
return finalMTU
}
// connectToSignal creates Signal Service client and established a connection
func connectToSignal(ctx context.Context, wtConfig *mgmProto.NetbirdConfig, ourPrivateKey wgtypes.Key) (*signal.GrpcClient, error) {
var sigTLSEnabled bool

View File

@@ -0,0 +1,201 @@
package config
import (
"errors"
"fmt"
"net"
"net/netip"
"net/url"
"strings"
log "github.com/sirupsen/logrus"
"github.com/netbirdio/netbird/shared/management/domain"
mgmProto "github.com/netbirdio/netbird/shared/management/proto"
)
var (
ErrEmptyURL = errors.New("empty URL")
ErrEmptyHost = errors.New("empty host")
ErrIPNotAllowed = errors.New("IP address not allowed")
)
// ServerDomains represents the management server domains extracted from NetBird configuration
type ServerDomains struct {
Signal domain.Domain
Relay []domain.Domain
Flow domain.Domain
Stuns []domain.Domain
Turns []domain.Domain
}
// ExtractFromNetbirdConfig extracts domain information from NetBird protobuf configuration
func ExtractFromNetbirdConfig(config *mgmProto.NetbirdConfig) ServerDomains {
if config == nil {
return ServerDomains{}
}
domains := ServerDomains{}
domains.Signal = extractSignalDomain(config)
domains.Relay = extractRelayDomains(config)
domains.Flow = extractFlowDomain(config)
domains.Stuns = extractStunDomains(config)
domains.Turns = extractTurnDomains(config)
return domains
}
// ExtractValidDomain extracts a valid domain from a URL, filtering out IP addresses
func ExtractValidDomain(rawURL string) (domain.Domain, error) {
if rawURL == "" {
return "", ErrEmptyURL
}
parsedURL, err := url.Parse(rawURL)
if err == nil {
if domain, err := extractFromParsedURL(parsedURL); err != nil || domain != "" {
return domain, err
}
}
return extractFromRawString(rawURL)
}
// extractFromParsedURL handles domain extraction from successfully parsed URLs
func extractFromParsedURL(parsedURL *url.URL) (domain.Domain, error) {
if parsedURL.Hostname() != "" {
return extractDomainFromHost(parsedURL.Hostname())
}
if parsedURL.Opaque == "" || parsedURL.Scheme == "" {
return "", nil
}
// Handle URLs with opaque content (e.g., stun:host:port)
if strings.Contains(parsedURL.Scheme, ".") {
// This is likely "domain.com:port" being parsed as scheme:opaque
reconstructed := parsedURL.Scheme + ":" + parsedURL.Opaque
if host, _, err := net.SplitHostPort(reconstructed); err == nil {
return extractDomainFromHost(host)
}
return extractDomainFromHost(parsedURL.Scheme)
}
// Valid scheme with opaque content (e.g., stun:host:port)
host := parsedURL.Opaque
if queryIndex := strings.Index(host, "?"); queryIndex > 0 {
host = host[:queryIndex]
}
if hostOnly, _, err := net.SplitHostPort(host); err == nil {
return extractDomainFromHost(hostOnly)
}
return extractDomainFromHost(host)
}
// extractFromRawString handles domain extraction when URL parsing fails or returns no results
func extractFromRawString(rawURL string) (domain.Domain, error) {
if host, _, err := net.SplitHostPort(rawURL); err == nil {
return extractDomainFromHost(host)
}
return extractDomainFromHost(rawURL)
}
// extractDomainFromHost extracts domain from a host string, filtering out IP addresses
func extractDomainFromHost(host string) (domain.Domain, error) {
if host == "" {
return "", ErrEmptyHost
}
if _, err := netip.ParseAddr(host); err == nil {
return "", fmt.Errorf("%w: %s", ErrIPNotAllowed, host)
}
d, err := domain.FromString(host)
if err != nil {
return "", fmt.Errorf("invalid domain: %v", err)
}
return d, nil
}
// extractSingleDomain extracts a single domain from a URL with error logging
func extractSingleDomain(url, serviceType string) domain.Domain {
if url == "" {
return ""
}
d, err := ExtractValidDomain(url)
if err != nil {
log.Debugf("Skipping %s: %v", serviceType, err)
return ""
}
return d
}
// extractMultipleDomains extracts multiple domains from URLs with error logging
func extractMultipleDomains(urls []string, serviceType string) []domain.Domain {
var domains []domain.Domain
for _, url := range urls {
if url == "" {
continue
}
d, err := ExtractValidDomain(url)
if err != nil {
log.Debugf("Skipping %s: %v", serviceType, err)
continue
}
domains = append(domains, d)
}
return domains
}
// extractSignalDomain extracts the signal domain from NetBird configuration.
func extractSignalDomain(config *mgmProto.NetbirdConfig) domain.Domain {
if config.Signal != nil {
return extractSingleDomain(config.Signal.Uri, "signal")
}
return ""
}
// extractRelayDomains extracts relay server domains from NetBird configuration.
func extractRelayDomains(config *mgmProto.NetbirdConfig) []domain.Domain {
if config.Relay != nil {
return extractMultipleDomains(config.Relay.Urls, "relay")
}
return nil
}
// extractFlowDomain extracts the traffic flow domain from NetBird configuration.
func extractFlowDomain(config *mgmProto.NetbirdConfig) domain.Domain {
if config.Flow != nil {
return extractSingleDomain(config.Flow.Url, "flow")
}
return ""
}
// extractStunDomains extracts STUN server domains from NetBird configuration.
func extractStunDomains(config *mgmProto.NetbirdConfig) []domain.Domain {
var urls []string
for _, stun := range config.Stuns {
if stun != nil && stun.Uri != "" {
urls = append(urls, stun.Uri)
}
}
return extractMultipleDomains(urls, "STUN")
}
// extractTurnDomains extracts TURN server domains from NetBird configuration.
func extractTurnDomains(config *mgmProto.NetbirdConfig) []domain.Domain {
var urls []string
for _, turn := range config.Turns {
if turn != nil && turn.HostConfig != nil && turn.HostConfig.Uri != "" {
urls = append(urls, turn.HostConfig.Uri)
}
}
return extractMultipleDomains(urls, "TURN")
}

View File

@@ -0,0 +1,213 @@
package config
import (
"testing"
"github.com/stretchr/testify/assert"
)
func TestExtractValidDomain(t *testing.T) {
tests := []struct {
name string
url string
expected string
expectError bool
}{
{
name: "HTTPS URL with port",
url: "https://api.netbird.io:443",
expected: "api.netbird.io",
},
{
name: "HTTP URL without port",
url: "http://signal.example.com",
expected: "signal.example.com",
},
{
name: "Host with port (no scheme)",
url: "signal.netbird.io:443",
expected: "signal.netbird.io",
},
{
name: "STUN URL",
url: "stun:stun.netbird.io:443",
expected: "stun.netbird.io",
},
{
name: "STUN URL with different port",
url: "stun:stun.netbird.io:5555",
expected: "stun.netbird.io",
},
{
name: "TURNS URL with query params",
url: "turns:turn.netbird.io:443?transport=tcp",
expected: "turn.netbird.io",
},
{
name: "TURN URL",
url: "turn:turn.example.com:3478",
expected: "turn.example.com",
},
{
name: "REL URL",
url: "rel://relay.example.com:443",
expected: "relay.example.com",
},
{
name: "RELS URL",
url: "rels://relay.netbird.io:443",
expected: "relay.netbird.io",
},
{
name: "Raw hostname",
url: "example.org",
expected: "example.org",
},
{
name: "IP address should be rejected",
url: "192.168.1.1",
expectError: true,
},
{
name: "IP address with port should be rejected",
url: "192.168.1.1:443",
expectError: true,
},
{
name: "IPv6 address should be rejected",
url: "2001:db8::1",
expectError: true,
},
{
name: "HTTP URL with IPv4 should be rejected",
url: "http://192.168.1.1:8080",
expectError: true,
},
{
name: "HTTPS URL with IPv4 should be rejected",
url: "https://10.0.0.1:443",
expectError: true,
},
{
name: "STUN URL with IPv4 should be rejected",
url: "stun:192.168.1.1:3478",
expectError: true,
},
{
name: "TURN URL with IPv4 should be rejected",
url: "turn:10.0.0.1:3478",
expectError: true,
},
{
name: "TURNS URL with IPv4 should be rejected",
url: "turns:172.16.0.1:5349",
expectError: true,
},
{
name: "HTTP URL with IPv6 should be rejected",
url: "http://[2001:db8::1]:8080",
expectError: true,
},
{
name: "HTTPS URL with IPv6 should be rejected",
url: "https://[::1]:443",
expectError: true,
},
{
name: "STUN URL with IPv6 should be rejected",
url: "stun:[2001:db8::1]:3478",
expectError: true,
},
{
name: "IPv6 with port should be rejected",
url: "[2001:db8::1]:443",
expectError: true,
},
{
name: "Localhost IPv4 should be rejected",
url: "127.0.0.1:8080",
expectError: true,
},
{
name: "Localhost IPv6 should be rejected",
url: "[::1]:443",
expectError: true,
},
{
name: "REL URL with IPv4 should be rejected",
url: "rel://192.168.1.1:443",
expectError: true,
},
{
name: "RELS URL with IPv4 should be rejected",
url: "rels://10.0.0.1:443",
expectError: true,
},
{
name: "Empty URL",
url: "",
expectError: true,
},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
result, err := ExtractValidDomain(tt.url)
if tt.expectError {
assert.Error(t, err, "Expected error for URL: %s", tt.url)
} else {
assert.NoError(t, err, "Unexpected error for URL: %s", tt.url)
assert.Equal(t, tt.expected, result.SafeString(), "Domain mismatch for URL: %s", tt.url)
}
})
}
}
func TestExtractDomainFromHost(t *testing.T) {
tests := []struct {
name string
host string
expected string
expectError bool
}{
{
name: "Valid domain",
host: "example.com",
expected: "example.com",
},
{
name: "Subdomain",
host: "api.example.com",
expected: "api.example.com",
},
{
name: "IPv4 address",
host: "192.168.1.1",
expectError: true,
},
{
name: "IPv6 address",
host: "2001:db8::1",
expectError: true,
},
{
name: "Empty host",
host: "",
expectError: true,
},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
result, err := extractDomainFromHost(tt.host)
if tt.expectError {
assert.Error(t, err, "Expected error for host: %s", tt.host)
} else {
assert.NoError(t, err, "Unexpected error for host: %s", tt.host)
assert.Equal(t, tt.expected, result.SafeString(), "Domain mismatch for host: %s", tt.host)
}
})
}
}

View File

@@ -176,4 +176,3 @@ nameserver 192.168.0.1
t.Errorf("unexpected resolv.conf content: %v", cfg)
}
}

View File

@@ -11,11 +11,12 @@ import (
)
const (
PriorityLocal = 100
PriorityDNSRoute = 75
PriorityUpstream = 50
PriorityDefault = 1
PriorityFallback = -100
PriorityMgmtCache = 150
PriorityLocal = 100
PriorityDNSRoute = 75
PriorityUpstream = 50
PriorityDefault = 1
PriorityFallback = -100
)
type SubdomainMatcher interface {
@@ -182,7 +183,10 @@ func (c *HandlerChain) ServeDNS(w dns.ResponseWriter, r *dns.Msg) {
// If handler wants to continue, try next handler
if chainWriter.shouldContinue {
log.Tracef("handler requested continue to next handler for domain=%s", qname)
// 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

View File

@@ -166,9 +166,10 @@ func (s *systemConfigurator) removeKeyFromSystemConfig(key string) error {
func (s *systemConfigurator) addLocalDNS() error {
if !s.systemDNSSettings.ServerIP.IsValid() || len(s.systemDNSSettings.Domains) == 0 {
err := s.recordSystemDNSSettings(true)
log.Errorf("Unable to get system DNS configuration")
return err
if err := s.recordSystemDNSSettings(true); err != nil {
log.Errorf("Unable to get system DNS configuration")
return fmt.Errorf("recordSystemDNSSettings(): %w", err)
}
}
localKey := getKeyWithInput(netbirdDNSStateKeyFormat, localSuffix)
if s.systemDNSSettings.ServerIP.IsValid() && len(s.systemDNSSettings.Domains) != 0 {

View File

@@ -64,9 +64,10 @@ const (
)
type registryConfigurator struct {
guid string
routingAll bool
gpo bool
guid string
routingAll bool
gpo bool
nrptEntryCount int
}
func newHostManager(wgInterface WGIface) (*registryConfigurator, error) {
@@ -177,7 +178,11 @@ func (r *registryConfigurator) applyDNSConfig(config HostDNSConfig, stateManager
log.Infof("removed %s as main DNS forwarder for this peer", config.ServerIP)
}
if err := stateManager.UpdateState(&ShutdownState{Guid: r.guid, GPO: r.gpo}); err != nil {
if err := stateManager.UpdateState(&ShutdownState{
Guid: r.guid,
GPO: r.gpo,
NRPTEntryCount: r.nrptEntryCount,
}); err != nil {
log.Errorf("failed to update shutdown state: %s", err)
}
@@ -193,13 +198,24 @@ func (r *registryConfigurator) applyDNSConfig(config HostDNSConfig, stateManager
}
if len(matchDomains) != 0 {
if err := r.addDNSMatchPolicy(matchDomains, config.ServerIP); err != nil {
count, err := r.addDNSMatchPolicy(matchDomains, config.ServerIP)
if err != nil {
return fmt.Errorf("add dns match policy: %w", err)
}
r.nrptEntryCount = count
} else {
if err := r.removeDNSMatchPolicies(); err != nil {
return fmt.Errorf("remove dns match policies: %w", err)
}
r.nrptEntryCount = 0
}
if err := stateManager.UpdateState(&ShutdownState{
Guid: r.guid,
GPO: r.gpo,
NRPTEntryCount: r.nrptEntryCount,
}); err != nil {
log.Errorf("failed to update shutdown state: %s", err)
}
if err := r.updateSearchDomains(searchDomains); err != nil {
@@ -220,28 +236,36 @@ func (r *registryConfigurator) addDNSSetupForAll(ip netip.Addr) error {
return nil
}
func (r *registryConfigurator) addDNSMatchPolicy(domains []string, ip netip.Addr) error {
func (r *registryConfigurator) addDNSMatchPolicy(domains []string, ip netip.Addr) (int, error) {
// if the gpo key is present, we need to put our DNS settings there, otherwise our config might be ignored
// see https://learn.microsoft.com/en-us/openspecs/windows_protocols/ms-gpnrpt/8cc31cb9-20cb-4140-9e85-3e08703b4745
if r.gpo {
if err := r.configureDNSPolicy(gpoDnsPolicyConfigMatchPath, domains, ip); err != nil {
return fmt.Errorf("configure GPO DNS policy: %w", err)
for i, domain := range domains {
singleDomain := []string{domain}
if err := r.configureDNSPolicy(dnsPolicyConfigMatchPath, singleDomain, ip); err != nil {
return i, fmt.Errorf("configure DNS Local policy for domain %s: %w", domain, err)
}
if r.gpo {
if err := r.configureDNSPolicy(gpoDnsPolicyConfigMatchPath, singleDomain, ip); err != nil {
return i, fmt.Errorf("configure gpo DNS policy: %w", err)
}
}
log.Debugf("added NRPT entry for domain: %s", domain)
}
if r.gpo {
if err := refreshGroupPolicy(); err != nil {
log.Warnf("failed to refresh group policy: %v", err)
}
} else {
if err := r.configureDNSPolicy(dnsPolicyConfigMatchPath, domains, ip); err != nil {
return fmt.Errorf("configure local DNS policy: %w", err)
}
}
log.Infof("added %d match domains. Domain list: %s", len(domains), domains)
return nil
log.Infof("added %d separate NRPT entries. Domain list: %s", len(domains), domains)
return len(domains), nil
}
// configureDNSPolicy handles the actual configuration of a DNS policy at the specified path
func (r *registryConfigurator) configureDNSPolicy(policyPath string, domains []string, ip netip.Addr) error {
if err := removeRegistryKeyFromDNSPolicyConfig(policyPath); err != nil {
return fmt.Errorf("remove existing dns policy: %w", err)
@@ -374,12 +398,27 @@ func (r *registryConfigurator) restoreHostDNS() error {
func (r *registryConfigurator) removeDNSMatchPolicies() error {
var merr *multierror.Error
// Try to remove the base entries (for backward compatibility)
if err := removeRegistryKeyFromDNSPolicyConfig(dnsPolicyConfigMatchPath); err != nil {
merr = multierror.Append(merr, fmt.Errorf("remove local registry key: %w", err))
merr = multierror.Append(merr, fmt.Errorf("remove local base entry: %w", err))
}
if err := removeRegistryKeyFromDNSPolicyConfig(gpoDnsPolicyConfigMatchPath); err != nil {
merr = multierror.Append(merr, fmt.Errorf("remove GPO registry key: %w", err))
merr = multierror.Append(merr, fmt.Errorf("remove GPO base entry: %w", err))
}
for i := 0; i < r.nrptEntryCount; i++ {
localPath := fmt.Sprintf("%s-%d", dnsPolicyConfigMatchPath, i)
gpoPath := fmt.Sprintf("%s-%d", gpoDnsPolicyConfigMatchPath, i)
if err := removeRegistryKeyFromDNSPolicyConfig(localPath); err != nil {
merr = multierror.Append(merr, fmt.Errorf("remove local entry %d: %w", i, err))
}
if err := removeRegistryKeyFromDNSPolicyConfig(gpoPath); err != nil {
merr = multierror.Append(merr, fmt.Errorf("remove GPO entry %d: %w", i, err))
}
}
if err := refreshGroupPolicy(); err != nil {

View File

@@ -34,7 +34,7 @@ func (d *Resolver) MatchSubdomains() bool {
// String returns a string representation of the local resolver
func (d *Resolver) String() string {
return fmt.Sprintf("local resolver [%d records]", len(d.records))
return fmt.Sprintf("LocalResolver [%d records]", len(d.records))
}
func (d *Resolver) Stop() {}

View File

@@ -0,0 +1,360 @@
package mgmt
import (
"context"
"fmt"
"net"
"net/url"
"strings"
"sync"
"time"
"github.com/miekg/dns"
log "github.com/sirupsen/logrus"
dnsconfig "github.com/netbirdio/netbird/client/internal/dns/config"
"github.com/netbirdio/netbird/shared/management/domain"
)
const dnsTimeout = 5 * time.Second
// Resolver caches critical NetBird infrastructure domains
type Resolver struct {
records map[dns.Question][]dns.RR
mgmtDomain *domain.Domain
serverDomains *dnsconfig.ServerDomains
mutex sync.RWMutex
}
// NewResolver creates a new management domains cache resolver.
func NewResolver() *Resolver {
return &Resolver{
records: make(map[dns.Question][]dns.RR),
}
}
// String returns a string representation of the resolver.
func (m *Resolver) String() string {
return "MgmtCacheResolver"
}
// ServeDNS implements dns.Handler interface.
func (m *Resolver) ServeDNS(w dns.ResponseWriter, r *dns.Msg) {
if len(r.Question) == 0 {
m.continueToNext(w, r)
return
}
question := r.Question[0]
question.Name = strings.ToLower(dns.Fqdn(question.Name))
if question.Qtype != dns.TypeA && question.Qtype != dns.TypeAAAA {
m.continueToNext(w, r)
return
}
m.mutex.RLock()
records, found := m.records[question]
m.mutex.RUnlock()
if !found {
m.continueToNext(w, r)
return
}
resp := &dns.Msg{}
resp.SetReply(r)
resp.Authoritative = false
resp.RecursionAvailable = true
resp.Answer = append(resp.Answer, records...)
log.Debugf("serving %d cached records for domain=%s", len(resp.Answer), question.Name)
if err := w.WriteMsg(resp); err != nil {
log.Errorf("failed to write response: %v", err)
}
}
// MatchSubdomains returns false since this resolver only handles exact domain matches
// for NetBird infrastructure domains (signal, relay, flow, etc.), not their subdomains.
func (m *Resolver) MatchSubdomains() bool {
return false
}
// continueToNext signals the handler chain to continue to the next handler.
func (m *Resolver) continueToNext(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 {
log.Errorf("failed to write continue signal: %v", err)
}
}
// AddDomain manually adds a domain to cache by resolving it.
func (m *Resolver) AddDomain(ctx context.Context, d domain.Domain) error {
dnsName := strings.ToLower(dns.Fqdn(d.PunycodeString()))
ctx, cancel := context.WithTimeout(ctx, dnsTimeout)
defer cancel()
ips, err := net.DefaultResolver.LookupNetIP(ctx, "ip", d.PunycodeString())
if err != nil {
return fmt.Errorf("resolve domain %s: %w", d.SafeString(), err)
}
var aRecords, aaaaRecords []dns.RR
for _, ip := range ips {
if ip.Is4() {
rr := &dns.A{
Hdr: dns.RR_Header{
Name: dnsName,
Rrtype: dns.TypeA,
Class: dns.ClassINET,
Ttl: 300,
},
A: ip.AsSlice(),
}
aRecords = append(aRecords, rr)
} else if ip.Is6() {
rr := &dns.AAAA{
Hdr: dns.RR_Header{
Name: dnsName,
Rrtype: dns.TypeAAAA,
Class: dns.ClassINET,
Ttl: 300,
},
AAAA: ip.AsSlice(),
}
aaaaRecords = append(aaaaRecords, rr)
}
}
m.mutex.Lock()
if len(aRecords) > 0 {
aQuestion := dns.Question{
Name: dnsName,
Qtype: dns.TypeA,
Qclass: dns.ClassINET,
}
m.records[aQuestion] = aRecords
}
if len(aaaaRecords) > 0 {
aaaaQuestion := dns.Question{
Name: dnsName,
Qtype: dns.TypeAAAA,
Qclass: dns.ClassINET,
}
m.records[aaaaQuestion] = aaaaRecords
}
m.mutex.Unlock()
log.Debugf("added domain=%s with %d A records and %d AAAA records",
d.SafeString(), len(aRecords), len(aaaaRecords))
return nil
}
// PopulateFromConfig extracts and caches domains from the client configuration.
func (m *Resolver) PopulateFromConfig(ctx context.Context, mgmtURL *url.URL) error {
if mgmtURL == nil {
return nil
}
d, err := dnsconfig.ExtractValidDomain(mgmtURL.String())
if err != nil {
return fmt.Errorf("extract domain from URL: %w", err)
}
m.mutex.Lock()
m.mgmtDomain = &d
m.mutex.Unlock()
if err := m.AddDomain(ctx, d); err != nil {
return fmt.Errorf("add domain: %w", err)
}
return nil
}
// RemoveDomain removes a domain from the cache.
func (m *Resolver) RemoveDomain(d domain.Domain) error {
dnsName := strings.ToLower(dns.Fqdn(d.PunycodeString()))
m.mutex.Lock()
defer m.mutex.Unlock()
aQuestion := dns.Question{
Name: dnsName,
Qtype: dns.TypeA,
Qclass: dns.ClassINET,
}
delete(m.records, aQuestion)
aaaaQuestion := dns.Question{
Name: dnsName,
Qtype: dns.TypeAAAA,
Qclass: dns.ClassINET,
}
delete(m.records, aaaaQuestion)
log.Debugf("removed domain=%s from cache", d.SafeString())
return nil
}
// GetCachedDomains returns a list of all cached domains.
func (m *Resolver) GetCachedDomains() domain.List {
m.mutex.RLock()
defer m.mutex.RUnlock()
domainSet := make(map[domain.Domain]struct{})
for question := range m.records {
domainName := strings.TrimSuffix(question.Name, ".")
domainSet[domain.Domain(domainName)] = struct{}{}
}
domains := make(domain.List, 0, len(domainSet))
for d := range domainSet {
domains = append(domains, d)
}
return domains
}
// UpdateFromServerDomains updates the cache with server domains from network configuration.
// It merges new domains with existing ones, replacing entire domain types when updated.
// Empty updates are ignored to prevent clearing infrastructure domains during partial updates.
func (m *Resolver) UpdateFromServerDomains(ctx context.Context, serverDomains dnsconfig.ServerDomains) (domain.List, error) {
newDomains := m.extractDomainsFromServerDomains(serverDomains)
var removedDomains domain.List
if len(newDomains) > 0 {
m.mutex.Lock()
if m.serverDomains == nil {
m.serverDomains = &dnsconfig.ServerDomains{}
}
updatedServerDomains := m.mergeServerDomains(*m.serverDomains, serverDomains)
m.serverDomains = &updatedServerDomains
m.mutex.Unlock()
allDomains := m.extractDomainsFromServerDomains(updatedServerDomains)
currentDomains := m.GetCachedDomains()
removedDomains = m.removeStaleDomains(currentDomains, allDomains)
}
m.addNewDomains(ctx, newDomains)
return removedDomains, nil
}
// removeStaleDomains removes cached domains not present in the target domain list.
// Management domains are preserved and never removed during server domain updates.
func (m *Resolver) removeStaleDomains(currentDomains, newDomains domain.List) domain.List {
var removedDomains domain.List
for _, currentDomain := range currentDomains {
if m.isDomainInList(currentDomain, newDomains) {
continue
}
if m.isManagementDomain(currentDomain) {
continue
}
removedDomains = append(removedDomains, currentDomain)
if err := m.RemoveDomain(currentDomain); err != nil {
log.Warnf("failed to remove domain=%s: %v", currentDomain.SafeString(), err)
}
}
return removedDomains
}
// mergeServerDomains merges new server domains with existing ones.
// When a domain type is provided in the new domains, it completely replaces that type.
func (m *Resolver) mergeServerDomains(existing, incoming dnsconfig.ServerDomains) dnsconfig.ServerDomains {
merged := existing
if incoming.Signal != "" {
merged.Signal = incoming.Signal
}
if len(incoming.Relay) > 0 {
merged.Relay = incoming.Relay
}
if incoming.Flow != "" {
merged.Flow = incoming.Flow
}
if len(incoming.Stuns) > 0 {
merged.Stuns = incoming.Stuns
}
if len(incoming.Turns) > 0 {
merged.Turns = incoming.Turns
}
return merged
}
// isDomainInList checks if domain exists in the list
func (m *Resolver) isDomainInList(domain domain.Domain, list domain.List) bool {
for _, d := range list {
if domain.SafeString() == d.SafeString() {
return true
}
}
return false
}
// isManagementDomain checks if domain is the protected management domain
func (m *Resolver) isManagementDomain(domain domain.Domain) bool {
m.mutex.RLock()
defer m.mutex.RUnlock()
return m.mgmtDomain != nil && domain == *m.mgmtDomain
}
// addNewDomains resolves and caches all domains from the update
func (m *Resolver) addNewDomains(ctx context.Context, newDomains domain.List) {
for _, newDomain := range newDomains {
if err := m.AddDomain(ctx, newDomain); err != nil {
log.Warnf("failed to add/update domain=%s: %v", newDomain.SafeString(), err)
} else {
log.Debugf("added/updated management cache domain=%s", newDomain.SafeString())
}
}
}
func (m *Resolver) extractDomainsFromServerDomains(serverDomains dnsconfig.ServerDomains) domain.List {
var domains domain.List
if serverDomains.Signal != "" {
domains = append(domains, serverDomains.Signal)
}
for _, relay := range serverDomains.Relay {
if relay != "" {
domains = append(domains, relay)
}
}
if serverDomains.Flow != "" {
domains = append(domains, serverDomains.Flow)
}
for _, stun := range serverDomains.Stuns {
if stun != "" {
domains = append(domains, stun)
}
}
for _, turn := range serverDomains.Turns {
if turn != "" {
domains = append(domains, turn)
}
}
return domains
}

View File

@@ -0,0 +1,416 @@
package mgmt
import (
"context"
"fmt"
"net/url"
"strings"
"testing"
"github.com/miekg/dns"
"github.com/stretchr/testify/assert"
dnsconfig "github.com/netbirdio/netbird/client/internal/dns/config"
"github.com/netbirdio/netbird/client/internal/dns/test"
"github.com/netbirdio/netbird/shared/management/domain"
)
func TestResolver_NewResolver(t *testing.T) {
resolver := NewResolver()
assert.NotNil(t, resolver)
assert.NotNil(t, resolver.records)
assert.False(t, resolver.MatchSubdomains())
}
func TestResolver_ExtractDomainFromURL(t *testing.T) {
tests := []struct {
name string
urlStr string
expectedDom string
expectError bool
}{
{
name: "HTTPS URL with port",
urlStr: "https://api.netbird.io:443",
expectedDom: "api.netbird.io",
expectError: false,
},
{
name: "HTTP URL without port",
urlStr: "http://signal.example.com",
expectedDom: "signal.example.com",
expectError: false,
},
{
name: "URL with path",
urlStr: "https://relay.netbird.io/status",
expectedDom: "relay.netbird.io",
expectError: false,
},
{
name: "Invalid URL",
urlStr: "not-a-valid-url",
expectedDom: "not-a-valid-url",
expectError: false,
},
{
name: "Empty URL",
urlStr: "",
expectedDom: "",
expectError: true,
},
{
name: "STUN URL",
urlStr: "stun:stun.example.com:3478",
expectedDom: "stun.example.com",
expectError: false,
},
{
name: "TURN URL",
urlStr: "turn:turn.example.com:3478",
expectedDom: "turn.example.com",
expectError: false,
},
{
name: "REL URL",
urlStr: "rel://relay.example.com:443",
expectedDom: "relay.example.com",
expectError: false,
},
{
name: "RELS URL",
urlStr: "rels://relay.example.com:443",
expectedDom: "relay.example.com",
expectError: false,
},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
var parsedURL *url.URL
var err error
if tt.urlStr != "" {
parsedURL, err = url.Parse(tt.urlStr)
if err != nil && !tt.expectError {
t.Fatalf("Failed to parse URL: %v", err)
}
}
domain, err := extractDomainFromURL(parsedURL)
if tt.expectError {
assert.Error(t, err)
} else {
assert.NoError(t, err)
assert.Equal(t, tt.expectedDom, domain.SafeString())
}
})
}
}
func TestResolver_PopulateFromConfig(t *testing.T) {
ctx, cancel := context.WithCancel(context.Background())
defer cancel()
resolver := NewResolver()
// Test with IP address - should return error since IP addresses are rejected
mgmtURL, _ := url.Parse("https://127.0.0.1")
err := resolver.PopulateFromConfig(ctx, mgmtURL)
assert.Error(t, err)
assert.ErrorIs(t, err, dnsconfig.ErrIPNotAllowed)
// No domains should be cached when using IP addresses
domains := resolver.GetCachedDomains()
assert.Equal(t, 0, len(domains), "No domains should be cached when using IP addresses")
}
func TestResolver_ServeDNS(t *testing.T) {
resolver := NewResolver()
ctx := context.Background()
// Add a test domain to the cache - use example.org which is reserved for testing
testDomain, err := domain.FromString("example.org")
if err != nil {
t.Fatalf("Failed to create domain: %v", err)
}
err = resolver.AddDomain(ctx, testDomain)
if err != nil {
t.Skipf("Skipping test due to DNS resolution failure: %v", err)
}
// Test A record query for cached domain
t.Run("Cached domain A record", func(t *testing.T) {
var capturedMsg *dns.Msg
mockWriter := &test.MockResponseWriter{
WriteMsgFunc: func(m *dns.Msg) error {
capturedMsg = m
return nil
},
}
req := new(dns.Msg)
req.SetQuestion("example.org.", dns.TypeA)
resolver.ServeDNS(mockWriter, req)
assert.NotNil(t, capturedMsg)
assert.Equal(t, dns.RcodeSuccess, capturedMsg.Rcode)
assert.True(t, len(capturedMsg.Answer) > 0, "Should have at least one answer")
})
// Test uncached domain signals to continue to next handler
t.Run("Uncached domain signals continue to next handler", func(t *testing.T) {
var capturedMsg *dns.Msg
mockWriter := &test.MockResponseWriter{
WriteMsgFunc: func(m *dns.Msg) error {
capturedMsg = m
return nil
},
}
req := new(dns.Msg)
req.SetQuestion("unknown.example.com.", dns.TypeA)
resolver.ServeDNS(mockWriter, req)
assert.NotNil(t, capturedMsg)
assert.Equal(t, dns.RcodeNameError, capturedMsg.Rcode)
// Zero flag set to true signals the handler chain to continue to next handler
assert.True(t, capturedMsg.MsgHdr.Zero, "Zero flag should be set to signal continuation to next handler")
assert.Empty(t, capturedMsg.Answer, "Should have no answers for uncached domain")
})
// Test that subdomains of cached domains are NOT resolved
t.Run("Subdomains of cached domains are not resolved", func(t *testing.T) {
var capturedMsg *dns.Msg
mockWriter := &test.MockResponseWriter{
WriteMsgFunc: func(m *dns.Msg) error {
capturedMsg = m
return nil
},
}
// Query for a subdomain of our cached domain
req := new(dns.Msg)
req.SetQuestion("sub.example.org.", dns.TypeA)
resolver.ServeDNS(mockWriter, req)
assert.NotNil(t, capturedMsg)
assert.Equal(t, dns.RcodeNameError, capturedMsg.Rcode)
assert.True(t, capturedMsg.MsgHdr.Zero, "Should signal continuation to next handler for subdomains")
assert.Empty(t, capturedMsg.Answer, "Should have no answers for subdomains")
})
// Test case-insensitive matching
t.Run("Case-insensitive domain matching", func(t *testing.T) {
var capturedMsg *dns.Msg
mockWriter := &test.MockResponseWriter{
WriteMsgFunc: func(m *dns.Msg) error {
capturedMsg = m
return nil
},
}
// Query with different casing
req := new(dns.Msg)
req.SetQuestion("EXAMPLE.ORG.", dns.TypeA)
resolver.ServeDNS(mockWriter, req)
assert.NotNil(t, capturedMsg)
assert.Equal(t, dns.RcodeSuccess, capturedMsg.Rcode)
assert.True(t, len(capturedMsg.Answer) > 0, "Should resolve regardless of case")
})
}
func TestResolver_GetCachedDomains(t *testing.T) {
resolver := NewResolver()
ctx := context.Background()
testDomain, err := domain.FromString("example.org")
if err != nil {
t.Fatalf("Failed to create domain: %v", err)
}
err = resolver.AddDomain(ctx, testDomain)
if err != nil {
t.Skipf("Skipping test due to DNS resolution failure: %v", err)
}
cachedDomains := resolver.GetCachedDomains()
assert.Equal(t, 1, len(cachedDomains), "Should return exactly one domain for single added domain")
assert.Equal(t, testDomain.SafeString(), cachedDomains[0].SafeString(), "Cached domain should match original")
assert.False(t, strings.HasSuffix(cachedDomains[0].PunycodeString(), "."), "Domain should not have trailing dot")
}
func TestResolver_ManagementDomainProtection(t *testing.T) {
resolver := NewResolver()
ctx := context.Background()
mgmtURL, _ := url.Parse("https://example.org")
err := resolver.PopulateFromConfig(ctx, mgmtURL)
if err != nil {
t.Skipf("Skipping test due to DNS resolution failure: %v", err)
}
initialDomains := resolver.GetCachedDomains()
if len(initialDomains) == 0 {
t.Skip("Management domain failed to resolve, skipping test")
}
assert.Equal(t, 1, len(initialDomains), "Should have management domain cached")
assert.Equal(t, "example.org", initialDomains[0].SafeString())
serverDomains := dnsconfig.ServerDomains{
Signal: "google.com",
Relay: []domain.Domain{"cloudflare.com"},
}
_, err = resolver.UpdateFromServerDomains(ctx, serverDomains)
if err != nil {
t.Logf("Server domains update failed: %v", err)
}
finalDomains := resolver.GetCachedDomains()
managementStillCached := false
for _, d := range finalDomains {
if d.SafeString() == "example.org" {
managementStillCached = true
break
}
}
assert.True(t, managementStillCached, "Management domain should never be removed")
}
// extractDomainFromURL extracts a domain from a URL - test helper function
func extractDomainFromURL(u *url.URL) (domain.Domain, error) {
if u == nil {
return "", fmt.Errorf("URL is nil")
}
return dnsconfig.ExtractValidDomain(u.String())
}
func TestResolver_EmptyUpdateDoesNotRemoveDomains(t *testing.T) {
resolver := NewResolver()
ctx := context.Background()
// Set up initial domains using resolvable domains
initialDomains := dnsconfig.ServerDomains{
Signal: "example.org",
Stuns: []domain.Domain{"google.com"},
Turns: []domain.Domain{"cloudflare.com"},
}
// Add initial domains
_, err := resolver.UpdateFromServerDomains(ctx, initialDomains)
if err != nil {
t.Skipf("Skipping test due to DNS resolution failure: %v", err)
}
// Verify domains were added
cachedDomains := resolver.GetCachedDomains()
assert.Len(t, cachedDomains, 3)
// Update with empty ServerDomains (simulating partial network map update)
emptyDomains := dnsconfig.ServerDomains{}
removedDomains, err := resolver.UpdateFromServerDomains(ctx, emptyDomains)
assert.NoError(t, err)
// Verify no domains were removed
assert.Len(t, removedDomains, 0, "No domains should be removed when update is empty")
// Verify all original domains are still cached
finalDomains := resolver.GetCachedDomains()
assert.Len(t, finalDomains, 3, "All original domains should still be cached")
}
func TestResolver_PartialUpdateReplacesOnlyUpdatedTypes(t *testing.T) {
resolver := NewResolver()
ctx := context.Background()
// Set up initial complete domains using resolvable domains
initialDomains := dnsconfig.ServerDomains{
Signal: "example.org",
Stuns: []domain.Domain{"google.com"},
Turns: []domain.Domain{"cloudflare.com"},
}
// Add initial domains
_, err := resolver.UpdateFromServerDomains(ctx, initialDomains)
if err != nil {
t.Skipf("Skipping test due to DNS resolution failure: %v", err)
}
assert.Len(t, resolver.GetCachedDomains(), 3)
// Update with partial ServerDomains (only signal domain - this should replace signal but preserve stun/turn)
partialDomains := dnsconfig.ServerDomains{
Signal: "github.com",
}
removedDomains, err := resolver.UpdateFromServerDomains(ctx, partialDomains)
if err != nil {
t.Skipf("Skipping test due to DNS resolution failure: %v", err)
}
// Should remove only the old signal domain
assert.Len(t, removedDomains, 1, "Should remove only the old signal domain")
assert.Equal(t, "example.org", removedDomains[0].SafeString())
finalDomains := resolver.GetCachedDomains()
assert.Len(t, finalDomains, 3, "Should have new signal plus preserved stun/turn domains")
domainStrings := make([]string, len(finalDomains))
for i, d := range finalDomains {
domainStrings[i] = d.SafeString()
}
assert.Contains(t, domainStrings, "github.com")
assert.Contains(t, domainStrings, "google.com")
assert.Contains(t, domainStrings, "cloudflare.com")
assert.NotContains(t, domainStrings, "example.org")
}
func TestResolver_PartialUpdateAddsNewTypePreservesExisting(t *testing.T) {
resolver := NewResolver()
ctx := context.Background()
// Set up initial complete domains using resolvable domains
initialDomains := dnsconfig.ServerDomains{
Signal: "example.org",
Stuns: []domain.Domain{"google.com"},
Turns: []domain.Domain{"cloudflare.com"},
}
// Add initial domains
_, err := resolver.UpdateFromServerDomains(ctx, initialDomains)
if err != nil {
t.Skipf("Skipping test due to DNS resolution failure: %v", err)
}
assert.Len(t, resolver.GetCachedDomains(), 3)
// Update with partial ServerDomains (only flow domain - new type, should preserve all existing)
partialDomains := dnsconfig.ServerDomains{
Flow: "github.com",
}
removedDomains, err := resolver.UpdateFromServerDomains(ctx, partialDomains)
if err != nil {
t.Skipf("Skipping test due to DNS resolution failure: %v", err)
}
assert.Len(t, removedDomains, 0, "Should not remove any domains when adding new type")
finalDomains := resolver.GetCachedDomains()
assert.Len(t, finalDomains, 4, "Should have all original domains plus new flow domain")
domainStrings := make([]string, len(finalDomains))
for i, d := range finalDomains {
domainStrings[i] = d.SafeString()
}
assert.Contains(t, domainStrings, "example.org")
assert.Contains(t, domainStrings, "google.com")
assert.Contains(t, domainStrings, "cloudflare.com")
assert.Contains(t, domainStrings, "github.com")
}

View File

@@ -3,20 +3,23 @@ package dns
import (
"fmt"
"net/netip"
"net/url"
"github.com/miekg/dns"
dnsconfig "github.com/netbirdio/netbird/client/internal/dns/config"
nbdns "github.com/netbirdio/netbird/dns"
"github.com/netbirdio/netbird/shared/management/domain"
)
// MockServer is the mock instance of a dns server
type MockServer struct {
InitializeFunc func() error
StopFunc func()
UpdateDNSServerFunc func(serial uint64, update nbdns.Config) error
RegisterHandlerFunc func(domain.List, dns.Handler, int)
DeregisterHandlerFunc func(domain.List, int)
InitializeFunc func() error
StopFunc func()
UpdateDNSServerFunc func(serial uint64, update nbdns.Config) error
RegisterHandlerFunc func(domain.List, dns.Handler, int)
DeregisterHandlerFunc func(domain.List, int)
UpdateServerConfigFunc func(domains dnsconfig.ServerDomains) error
}
func (m *MockServer) RegisterHandler(domains domain.List, handler dns.Handler, priority int) {
@@ -70,3 +73,14 @@ func (m *MockServer) SearchDomains() []string {
// ProbeAvailability mocks implementation of ProbeAvailability from the Server interface
func (m *MockServer) ProbeAvailability() {
}
func (m *MockServer) UpdateServerConfig(domains dnsconfig.ServerDomains) error {
if m.UpdateServerConfigFunc != nil {
return m.UpdateServerConfigFunc(domains)
}
return nil
}
func (m *MockServer) PopulateManagementDomain(mgmtURL *url.URL) error {
return nil
}

View File

@@ -5,6 +5,7 @@ import (
"errors"
"fmt"
"net/netip"
"net/url"
"runtime"
"strings"
"sync"
@@ -15,7 +16,9 @@ import (
"golang.org/x/exp/maps"
"github.com/netbirdio/netbird/client/iface/netstack"
dnsconfig "github.com/netbirdio/netbird/client/internal/dns/config"
"github.com/netbirdio/netbird/client/internal/dns/local"
"github.com/netbirdio/netbird/client/internal/dns/mgmt"
"github.com/netbirdio/netbird/client/internal/dns/types"
"github.com/netbirdio/netbird/client/internal/listener"
"github.com/netbirdio/netbird/client/internal/peer"
@@ -45,6 +48,8 @@ type Server interface {
OnUpdatedHostDNSServer(addrs []netip.AddrPort)
SearchDomains() []string
ProbeAvailability()
UpdateServerConfig(domains dnsconfig.ServerDomains) error
PopulateManagementDomain(mgmtURL *url.URL) error
}
type nsGroupsByDomain struct {
@@ -77,6 +82,8 @@ type DefaultServer struct {
handlerChain *HandlerChain
extraDomains map[domain.Domain]int
mgmtCacheResolver *mgmt.Resolver
// permanent related properties
permanent bool
hostsDNSHolder *hostsDNSHolder
@@ -104,18 +111,20 @@ type handlerWrapper struct {
type registeredHandlerMap map[types.HandlerID]handlerWrapper
// DefaultServerConfig holds configuration parameters for NewDefaultServer
type DefaultServerConfig struct {
WgInterface WGIface
CustomAddress string
StatusRecorder *peer.Status
StateManager *statemanager.Manager
DisableSys bool
}
// NewDefaultServer returns a new dns server
func NewDefaultServer(
ctx context.Context,
wgInterface WGIface,
customAddress string,
statusRecorder *peer.Status,
stateManager *statemanager.Manager,
disableSys bool,
) (*DefaultServer, error) {
func NewDefaultServer(ctx context.Context, config DefaultServerConfig) (*DefaultServer, error) {
var addrPort *netip.AddrPort
if customAddress != "" {
parsedAddrPort, err := netip.ParseAddrPort(customAddress)
if config.CustomAddress != "" {
parsedAddrPort, err := netip.ParseAddrPort(config.CustomAddress)
if err != nil {
return nil, fmt.Errorf("unable to parse the custom dns address, got error: %s", err)
}
@@ -123,13 +132,14 @@ func NewDefaultServer(
}
var dnsService service
if wgInterface.IsUserspaceBind() {
dnsService = NewServiceViaMemory(wgInterface)
if config.WgInterface.IsUserspaceBind() {
dnsService = NewServiceViaMemory(config.WgInterface)
} else {
dnsService = newServiceViaListener(wgInterface, addrPort)
dnsService = newServiceViaListener(config.WgInterface, addrPort)
}
return newDefaultServer(ctx, wgInterface, dnsService, statusRecorder, stateManager, disableSys), nil
server := newDefaultServer(ctx, config.WgInterface, dnsService, config.StatusRecorder, config.StateManager, config.DisableSys)
return server, nil
}
// NewDefaultServerPermanentUpstream returns a new dns server. It optimized for mobile systems
@@ -178,20 +188,24 @@ func newDefaultServer(
) *DefaultServer {
handlerChain := NewHandlerChain()
ctx, stop := context.WithCancel(ctx)
mgmtCacheResolver := mgmt.NewResolver()
defaultServer := &DefaultServer{
ctx: ctx,
ctxCancel: stop,
disableSys: disableSys,
service: dnsService,
handlerChain: handlerChain,
extraDomains: make(map[domain.Domain]int),
dnsMuxMap: make(registeredHandlerMap),
localResolver: local.NewResolver(),
wgInterface: wgInterface,
statusRecorder: statusRecorder,
stateManager: stateManager,
hostsDNSHolder: newHostsDNSHolder(),
hostManager: &noopHostConfigurator{},
ctx: ctx,
ctxCancel: stop,
disableSys: disableSys,
service: dnsService,
handlerChain: handlerChain,
extraDomains: make(map[domain.Domain]int),
dnsMuxMap: make(registeredHandlerMap),
localResolver: local.NewResolver(),
wgInterface: wgInterface,
statusRecorder: statusRecorder,
stateManager: stateManager,
hostsDNSHolder: newHostsDNSHolder(),
hostManager: &noopHostConfigurator{},
mgmtCacheResolver: mgmtCacheResolver,
}
// register with root zone, handler chain takes care of the routing
@@ -217,7 +231,7 @@ func (s *DefaultServer) RegisterHandler(domains domain.List, handler dns.Handler
}
func (s *DefaultServer) registerHandler(domains []string, handler dns.Handler, priority int) {
log.Debugf("registering handler %s with priority %d", handler, priority)
log.Debugf("registering handler %s with priority %d for %v", handler, priority, domains)
for _, domain := range domains {
if domain == "" {
@@ -246,7 +260,7 @@ func (s *DefaultServer) DeregisterHandler(domains domain.List, priority int) {
}
func (s *DefaultServer) deregisterHandler(domains []string, priority int) {
log.Debugf("deregistering handler %v with priority %d", domains, priority)
log.Debugf("deregistering handler with priority %d for %v", priority, domains)
for _, domain := range domains {
if domain == "" {
@@ -432,6 +446,29 @@ func (s *DefaultServer) ProbeAvailability() {
wg.Wait()
}
func (s *DefaultServer) UpdateServerConfig(domains dnsconfig.ServerDomains) error {
s.mux.Lock()
defer s.mux.Unlock()
if s.mgmtCacheResolver != nil {
removedDomains, err := s.mgmtCacheResolver.UpdateFromServerDomains(s.ctx, domains)
if err != nil {
return fmt.Errorf("update management cache resolver: %w", err)
}
if len(removedDomains) > 0 {
s.deregisterHandler(removedDomains.ToPunycodeList(), PriorityMgmtCache)
}
newDomains := s.mgmtCacheResolver.GetCachedDomains()
if len(newDomains) > 0 {
s.registerHandler(newDomains.ToPunycodeList(), s.mgmtCacheResolver, PriorityMgmtCache)
}
}
return nil
}
func (s *DefaultServer) applyConfiguration(update nbdns.Config) error {
// is the service should be Disabled, we stop the listener or fake resolver
if update.ServiceEnable {
@@ -695,6 +732,12 @@ func (s *DefaultServer) createHandlersForDomainGroup(domainGroup nsGroupsByDomai
ns.IP.String(), ns.NSType.String(), nbdns.UDPNameServerType.String())
continue
}
if ns.IP == s.service.RuntimeIP() {
log.Warnf("skipping nameserver %s as it matches our DNS server IP, preventing potential loop", ns.IP)
continue
}
handler.upstreamServers = append(handler.upstreamServers, ns.AddrPort())
}
@@ -955,3 +998,11 @@ func toZone(d domain.Domain) domain.Domain {
),
)
}
// PopulateManagementDomain populates the DNS cache with management domain
func (s *DefaultServer) PopulateManagementDomain(mgmtURL *url.URL) error {
if s.mgmtCacheResolver != nil {
return s.mgmtCacheResolver.PopulateFromConfig(s.ctx, mgmtURL)
}
return nil
}

View File

@@ -363,7 +363,13 @@ func TestUpdateDNSServer(t *testing.T) {
t.Log(err)
}
}()
dnsServer, err := NewDefaultServer(context.Background(), wgIface, "", peer.NewRecorder("mgm"), nil, false)
dnsServer, err := NewDefaultServer(context.Background(), DefaultServerConfig{
WgInterface: wgIface,
CustomAddress: "",
StatusRecorder: peer.NewRecorder("mgm"),
StateManager: nil,
DisableSys: false,
})
if err != nil {
t.Fatal(err)
}
@@ -473,7 +479,13 @@ func TestDNSFakeResolverHandleUpdates(t *testing.T) {
return
}
dnsServer, err := NewDefaultServer(context.Background(), wgIface, "", peer.NewRecorder("mgm"), nil, false)
dnsServer, err := NewDefaultServer(context.Background(), DefaultServerConfig{
WgInterface: wgIface,
CustomAddress: "",
StatusRecorder: peer.NewRecorder("mgm"),
StateManager: nil,
DisableSys: false,
})
if err != nil {
t.Errorf("create DNS server: %v", err)
return
@@ -575,7 +587,13 @@ func TestDNSServerStartStop(t *testing.T) {
for _, testCase := range testCases {
t.Run(testCase.name, func(t *testing.T) {
dnsServer, err := NewDefaultServer(context.Background(), &mocWGIface{}, testCase.addrPort, peer.NewRecorder("mgm"), nil, false)
dnsServer, err := NewDefaultServer(context.Background(), DefaultServerConfig{
WgInterface: &mocWGIface{},
CustomAddress: testCase.addrPort,
StatusRecorder: peer.NewRecorder("mgm"),
StateManager: nil,
DisableSys: false,
})
if err != nil {
t.Fatalf("%v", err)
}
@@ -2056,3 +2074,124 @@ func TestLocalResolverPriorityConstants(t *testing.T) {
assert.Equal(t, PriorityLocal, localMuxUpdates[0].priority, "Local handler should use PriorityLocal")
assert.Equal(t, "local.example.com", localMuxUpdates[0].domain)
}
func TestDNSLoopPrevention(t *testing.T) {
wgInterface := &mocWGIface{}
service := NewServiceViaMemory(wgInterface)
dnsServerIP := service.RuntimeIP()
server := &DefaultServer{
ctx: context.Background(),
wgInterface: wgInterface,
service: service,
localResolver: local.NewResolver(),
handlerChain: NewHandlerChain(),
hostManager: &noopHostConfigurator{},
dnsMuxMap: make(registeredHandlerMap),
}
tests := []struct {
name string
nsGroups []*nbdns.NameServerGroup
expectedHandlers int
expectedServers []netip.Addr
shouldFilterOwnIP bool
}{
{
name: "FilterOwnDNSServerIP",
nsGroups: []*nbdns.NameServerGroup{
{
Primary: true,
NameServers: []nbdns.NameServer{
{IP: netip.MustParseAddr("8.8.8.8"), NSType: nbdns.UDPNameServerType, Port: 53},
{IP: dnsServerIP, NSType: nbdns.UDPNameServerType, Port: 53},
{IP: netip.MustParseAddr("1.1.1.1"), NSType: nbdns.UDPNameServerType, Port: 53},
},
Domains: []string{},
},
},
expectedHandlers: 1,
expectedServers: []netip.Addr{netip.MustParseAddr("8.8.8.8"), netip.MustParseAddr("1.1.1.1")},
shouldFilterOwnIP: true,
},
{
name: "AllServersFiltered",
nsGroups: []*nbdns.NameServerGroup{
{
Primary: false,
NameServers: []nbdns.NameServer{
{IP: dnsServerIP, NSType: nbdns.UDPNameServerType, Port: 53},
},
Domains: []string{"example.com"},
},
},
expectedHandlers: 0,
expectedServers: []netip.Addr{},
shouldFilterOwnIP: true,
},
{
name: "MixedServersWithOwnIP",
nsGroups: []*nbdns.NameServerGroup{
{
Primary: false,
NameServers: []nbdns.NameServer{
{IP: netip.MustParseAddr("8.8.8.8"), NSType: nbdns.UDPNameServerType, Port: 53},
{IP: dnsServerIP, NSType: nbdns.UDPNameServerType, Port: 53},
{IP: netip.MustParseAddr("1.1.1.1"), NSType: nbdns.UDPNameServerType, Port: 53},
{IP: dnsServerIP, NSType: nbdns.UDPNameServerType, Port: 53}, // duplicate
},
Domains: []string{"test.com"},
},
},
expectedHandlers: 1,
expectedServers: []netip.Addr{netip.MustParseAddr("8.8.8.8"), netip.MustParseAddr("1.1.1.1")},
shouldFilterOwnIP: true,
},
{
name: "NoOwnIPInList",
nsGroups: []*nbdns.NameServerGroup{
{
Primary: true,
NameServers: []nbdns.NameServer{
{IP: netip.MustParseAddr("8.8.8.8"), NSType: nbdns.UDPNameServerType, Port: 53},
{IP: netip.MustParseAddr("1.1.1.1"), NSType: nbdns.UDPNameServerType, Port: 53},
},
Domains: []string{},
},
},
expectedHandlers: 1,
expectedServers: []netip.Addr{netip.MustParseAddr("8.8.8.8"), netip.MustParseAddr("1.1.1.1")},
shouldFilterOwnIP: false,
},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
muxUpdates, err := server.buildUpstreamHandlerUpdate(tt.nsGroups)
assert.NoError(t, err)
assert.Len(t, muxUpdates, tt.expectedHandlers)
if tt.expectedHandlers > 0 {
handler := muxUpdates[0].handler.(*upstreamResolver)
assert.Len(t, handler.upstreamServers, len(tt.expectedServers))
if tt.shouldFilterOwnIP {
for _, upstream := range handler.upstreamServers {
assert.NotEqual(t, dnsServerIP, upstream.Addr())
}
}
for _, expected := range tt.expectedServers {
found := false
for _, upstream := range handler.upstreamServers {
if upstream.Addr() == expected {
found = true
break
}
}
assert.True(t, found, "Expected server %s not found", expected)
}
}
})
}
}

View File

@@ -10,7 +10,7 @@ import (
"github.com/miekg/dns"
log "github.com/sirupsen/logrus"
nbnet "github.com/netbirdio/netbird/util/net"
nbnet "github.com/netbirdio/netbird/client/net"
)
type ServiceViaMemory struct {

View File

@@ -5,8 +5,9 @@ import (
)
type ShutdownState struct {
Guid string
GPO bool
Guid string
GPO bool
NRPTEntryCount int
}
func (s *ShutdownState) Name() string {
@@ -15,8 +16,9 @@ func (s *ShutdownState) Name() string {
func (s *ShutdownState) Cleanup() error {
manager := &registryConfigurator{
guid: s.Guid,
gpo: s.GPO,
guid: s.Guid,
gpo: s.GPO,
nrptEntryCount: s.NRPTEntryCount,
}
if err := manager.restoreUncleanShutdownDNS(); err != nil {

View File

@@ -26,10 +26,18 @@ import (
"github.com/netbirdio/netbird/client/proto"
)
const (
UpstreamTimeout = 15 * time.Second
var currentMTU uint16 = iface.DefaultMTU
func SetCurrentMTU(mtu uint16) {
currentMTU = mtu
}
const (
UpstreamTimeout = 4 * time.Second
// ClientTimeout is the timeout for the dns.Client.
// Set longer than UpstreamTimeout to ensure context timeout takes precedence
ClientTimeout = 5 * time.Second
failsTillDeact = int32(5)
reactivatePeriod = 30 * time.Second
probeTimeout = 2 * time.Second
)
@@ -52,9 +60,7 @@ type upstreamResolverBase struct {
upstreamServers []netip.AddrPort
domain string
disabled bool
failsCount atomic.Int32
successCount atomic.Int32
failsTillDeact int32
mutex sync.Mutex
reactivatePeriod time.Duration
upstreamTimeout time.Duration
@@ -73,14 +79,13 @@ func newUpstreamResolverBase(ctx context.Context, statusRecorder *peer.Status, d
domain: domain,
upstreamTimeout: UpstreamTimeout,
reactivatePeriod: reactivatePeriod,
failsTillDeact: failsTillDeact,
statusRecorder: statusRecorder,
}
}
// String returns a string representation of the upstream resolver
func (u *upstreamResolverBase) String() string {
return fmt.Sprintf("upstream %s", u.upstreamServers)
return fmt.Sprintf("Upstream %s", u.upstreamServers)
}
// ID returns the unique handler ID
@@ -110,58 +115,102 @@ func (u *upstreamResolverBase) Stop() {
func (u *upstreamResolverBase) ServeDNS(w dns.ResponseWriter, r *dns.Msg) {
requestID := GenerateRequestID()
logger := log.WithField("request_id", requestID)
var err error
defer func() {
u.checkUpstreamFails(err)
}()
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)
if u.ctx.Err() != nil {
logger.Tracef("%s has been stopped", u)
return
}
if u.tryUpstreamServers(w, r, logger) {
return
}
u.writeErrorResponse(w, r, logger)
}
func (u *upstreamResolverBase) prepareRequest(r *dns.Msg) {
if r.Extra == nil {
r.MsgHdr.AuthenticatedData = true
}
}
select {
case <-u.ctx.Done():
logger.Tracef("%s has been stopped", u)
return
default:
func (u *upstreamResolverBase) tryUpstreamServers(w dns.ResponseWriter, r *dns.Msg, logger *log.Entry) bool {
timeout := u.upstreamTimeout
if len(u.upstreamServers) > 1 {
maxTotal := 5 * time.Second
minPerUpstream := 2 * time.Second
scaledTimeout := maxTotal / time.Duration(len(u.upstreamServers))
if scaledTimeout > minPerUpstream {
timeout = scaledTimeout
} else {
timeout = minPerUpstream
}
}
for _, upstream := range u.upstreamServers {
var rm *dns.Msg
var t time.Duration
func() {
ctx, cancel := context.WithTimeout(u.ctx, u.upstreamTimeout)
defer cancel()
rm, t, err = u.upstreamClient.exchange(ctx, upstream.String(), r)
}()
if err != nil {
if errors.Is(err, context.DeadlineExceeded) || isTimeout(err) {
logger.Warnf("upstream %s timed out for question domain=%s", upstream, r.Question[0].Name)
continue
}
logger.Warnf("failed to query upstream %s for question domain=%s: %s", upstream, r.Question[0].Name, err)
continue
if u.queryUpstream(w, r, upstream, timeout, logger) {
return true
}
}
return false
}
if rm == nil || !rm.Response {
logger.Warnf("no response from upstream %s for question domain=%s", upstream, r.Question[0].Name)
continue
}
func (u *upstreamResolverBase) queryUpstream(w dns.ResponseWriter, r *dns.Msg, upstream netip.AddrPort, timeout time.Duration, logger *log.Entry) bool {
var rm *dns.Msg
var t time.Duration
var err error
u.successCount.Add(1)
logger.Tracef("took %s to query the upstream %s for question domain=%s", t, upstream, r.Question[0].Name)
var startTime time.Time
func() {
ctx, cancel := context.WithTimeout(u.ctx, timeout)
defer cancel()
startTime = time.Now()
rm, t, err = u.upstreamClient.exchange(ctx, upstream.String(), r)
}()
if err = w.WriteMsg(rm); err != nil {
logger.Errorf("failed to write DNS response for question domain=%s: %s", r.Question[0].Name, err)
}
// count the fails only if they happen sequentially
u.failsCount.Store(0)
if err != nil {
u.handleUpstreamError(err, upstream, r.Question[0].Name, startTime, timeout, logger)
return false
}
if rm == nil || !rm.Response {
logger.Warnf("no response from upstream %s for question domain=%s", upstream, r.Question[0].Name)
return false
}
return u.writeSuccessResponse(w, rm, upstream, r.Question[0].Name, t, logger)
}
func (u *upstreamResolverBase) handleUpstreamError(err error, upstream netip.AddrPort, domain string, startTime time.Time, timeout time.Duration, logger *log.Entry) {
if !errors.Is(err, context.DeadlineExceeded) && !isTimeout(err) {
logger.Warnf("failed to query upstream %s for question domain=%s: %s", upstream, domain, err)
return
}
u.failsCount.Add(1)
elapsed := time.Since(startTime)
timeoutMsg := fmt.Sprintf("upstream %s timed out for question domain=%s after %v (timeout=%v)", upstream, domain, elapsed.Truncate(time.Millisecond), timeout)
if peerInfo := u.debugUpstreamTimeout(upstream); peerInfo != "" {
timeoutMsg += " " + peerInfo
}
timeoutMsg += fmt.Sprintf(" - error: %v", err)
logger.Warnf(timeoutMsg)
}
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)
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
}
func (u *upstreamResolverBase) writeErrorResponse(w dns.ResponseWriter, r *dns.Msg, logger *log.Entry) {
logger.Errorf("all queries to the %s failed for question domain=%s", u, r.Question[0].Name)
m := new(dns.Msg)
@@ -171,41 +220,6 @@ func (u *upstreamResolverBase) ServeDNS(w dns.ResponseWriter, r *dns.Msg) {
}
}
// checkUpstreamFails counts fails and disables or enables upstream resolving
//
// If fails count is greater that failsTillDeact, upstream resolving
// will be disabled for reactivatePeriod, after that time period fails counter
// will be reset and upstream will be reactivated.
func (u *upstreamResolverBase) checkUpstreamFails(err error) {
u.mutex.Lock()
defer u.mutex.Unlock()
if u.failsCount.Load() < u.failsTillDeact || u.disabled {
return
}
select {
case <-u.ctx.Done():
return
default:
}
u.disable(err)
if u.statusRecorder == nil {
return
}
u.statusRecorder.PublishEvent(
proto.SystemEvent_WARNING,
proto.SystemEvent_DNS,
"All upstream servers failed (fail count exceeded)",
"Unable to reach one or more DNS servers. This might affect your ability to connect to some services.",
map[string]string{"upstreams": u.upstreamServersString()},
// TODO add domain meta
)
}
// ProbeAvailability tests all upstream servers simultaneously and
// disables the resolver if none work
func (u *upstreamResolverBase) ProbeAvailability() {
@@ -218,8 +232,8 @@ func (u *upstreamResolverBase) ProbeAvailability() {
default:
}
// avoid probe if upstreams could resolve at least one query and fails count is less than failsTillDeact
if u.successCount.Load() > 0 && u.failsCount.Load() < u.failsTillDeact {
// avoid probe if upstreams could resolve at least one query
if u.successCount.Load() > 0 {
return
}
@@ -306,7 +320,6 @@ func (u *upstreamResolverBase) waitUntilResponse() {
}
log.Infof("upstreams %s are responsive again. Adding them back to system", u.upstreamServersString())
u.failsCount.Store(0)
u.successCount.Add(1)
u.reactivate()
u.disabled = false
@@ -358,8 +371,8 @@ func (u *upstreamResolverBase) testNameserver(server netip.AddrPort, timeout tim
// If the passed context is nil, this will use Exchange instead of ExchangeContext.
func ExchangeWithFallback(ctx context.Context, client *dns.Client, r *dns.Msg, upstream string) (*dns.Msg, time.Duration, error) {
// MTU - ip + udp headers
// Note: this could be sent out on an interface that is not ours, but our MTU should always be lower.
client.UDPSize = iface.DefaultMTU - (60 + 8)
// Note: this could be sent out on an interface that is not ours, but higher MTU settings could break truncation handling.
client.UDPSize = uint16(currentMTU - (60 + 8))
var (
rm *dns.Msg
@@ -410,3 +423,80 @@ func GenerateRequestID() string {
}
return hex.EncodeToString(bytes)
}
// FormatPeerStatus formats peer connection status information for debugging DNS timeouts
func FormatPeerStatus(peerState *peer.State) string {
isConnected := peerState.ConnStatus == peer.StatusConnected
hasRecentHandshake := !peerState.LastWireguardHandshake.IsZero() &&
time.Since(peerState.LastWireguardHandshake) < 3*time.Minute
statusInfo := fmt.Sprintf("%s:%s", peerState.FQDN, peerState.IP)
switch {
case !isConnected:
statusInfo += " DISCONNECTED"
case !hasRecentHandshake:
statusInfo += " NO_RECENT_HANDSHAKE"
default:
statusInfo += " connected"
}
if !peerState.LastWireguardHandshake.IsZero() {
timeSinceHandshake := time.Since(peerState.LastWireguardHandshake)
statusInfo += fmt.Sprintf(" last_handshake=%v_ago", timeSinceHandshake.Truncate(time.Second))
} else {
statusInfo += " no_handshake"
}
if peerState.Relayed {
statusInfo += " via_relay"
}
if peerState.Latency > 0 {
statusInfo += fmt.Sprintf(" latency=%v", peerState.Latency)
}
return statusInfo
}
// findPeerForIP finds which peer handles the given IP address
func findPeerForIP(ip netip.Addr, statusRecorder *peer.Status) *peer.State {
if statusRecorder == nil {
return nil
}
fullStatus := statusRecorder.GetFullStatus()
var bestMatch *peer.State
var bestPrefixLen int
for _, peerState := range fullStatus.Peers {
routes := peerState.GetRoutes()
for route := range routes {
prefix, err := netip.ParsePrefix(route)
if err != nil {
continue
}
if prefix.Contains(ip) && prefix.Bits() > bestPrefixLen {
peerStateCopy := peerState
bestMatch = &peerStateCopy
bestPrefixLen = prefix.Bits()
}
}
}
return bestMatch
}
func (u *upstreamResolverBase) debugUpstreamTimeout(upstream netip.AddrPort) string {
if u.statusRecorder == nil {
return ""
}
peerInfo := findPeerForIP(upstream.Addr(), u.statusRecorder)
if peerInfo == nil {
return ""
}
return fmt.Sprintf("(routes through NetBird peer %s)", FormatPeerStatus(peerInfo))
}

View File

@@ -10,7 +10,7 @@ import (
"github.com/miekg/dns"
"github.com/netbirdio/netbird/client/internal/peer"
nbnet "github.com/netbirdio/netbird/util/net"
nbnet "github.com/netbirdio/netbird/client/net"
)
type upstreamResolver struct {
@@ -50,7 +50,9 @@ func (u *upstreamResolver) exchange(ctx context.Context, upstream string, r *dns
}
func (u *upstreamResolver) exchangeWithinVPN(ctx context.Context, upstream string, r *dns.Msg) (rm *dns.Msg, t time.Duration, err error) {
upstreamExchangeClient := &dns.Client{}
upstreamExchangeClient := &dns.Client{
Timeout: ClientTimeout,
}
return upstreamExchangeClient.ExchangeContext(ctx, r, upstream)
}
@@ -72,10 +74,11 @@ func (u *upstreamResolver) exchangeWithoutVPN(ctx context.Context, upstream stri
}
upstreamExchangeClient := &dns.Client{
Dialer: dialer,
Dialer: dialer,
Timeout: timeout,
}
return upstreamExchangeClient.Exchange(r, upstream)
return upstreamExchangeClient.ExchangeContext(ctx, r, upstream)
}
func (u *upstreamResolver) isLocalResolver(upstream string) bool {

View File

@@ -34,7 +34,10 @@ func newUpstreamResolver(
}
func (u *upstreamResolver) exchange(ctx context.Context, upstream string, r *dns.Msg) (rm *dns.Msg, t time.Duration, err error) {
return ExchangeWithFallback(ctx, &dns.Client{}, r, upstream)
client := &dns.Client{
Timeout: ClientTimeout,
}
return ExchangeWithFallback(ctx, client, r, upstream)
}
func GetClientPrivate(ip netip.Addr, interfaceName string, dialTimeout time.Duration) (*dns.Client, error) {

View File

@@ -47,7 +47,9 @@ func newUpstreamResolver(
}
func (u *upstreamResolverIOS) exchange(ctx context.Context, upstream string, r *dns.Msg) (rm *dns.Msg, t time.Duration, err error) {
client := &dns.Client{}
client := &dns.Client{
Timeout: ClientTimeout,
}
upstreamHost, _, err := net.SplitHostPort(upstream)
if err != nil {
return nil, 0, fmt.Errorf("error while parsing upstream host: %s", err)
@@ -110,7 +112,8 @@ func GetClientPrivate(ip netip.Addr, interfaceName string, dialTimeout time.Dura
},
}
client := &dns.Client{
Dialer: dialer,
Dialer: dialer,
Timeout: dialTimeout,
}
return client, nil
}

View File

@@ -124,29 +124,26 @@ func (c mockUpstreamResolver) exchange(_ context.Context, _ string, _ *dns.Msg)
}
func TestUpstreamResolver_DeactivationReactivation(t *testing.T) {
mockClient := &mockUpstreamResolver{
err: dns.ErrTime,
r: new(dns.Msg),
rtt: time.Millisecond,
}
resolver := &upstreamResolverBase{
ctx: context.TODO(),
upstreamClient: &mockUpstreamResolver{
err: nil,
r: new(dns.Msg),
rtt: time.Millisecond,
},
ctx: context.TODO(),
upstreamClient: mockClient,
upstreamTimeout: UpstreamTimeout,
reactivatePeriod: reactivatePeriod,
failsTillDeact: failsTillDeact,
reactivatePeriod: time.Microsecond * 100,
}
addrPort, _ := netip.ParseAddrPort("0.0.0.0:1") // Use valid port for parsing, test will still fail on connection
resolver.upstreamServers = []netip.AddrPort{netip.AddrPortFrom(addrPort.Addr().Unmap(), addrPort.Port())}
resolver.failsTillDeact = 0
resolver.reactivatePeriod = time.Microsecond * 100
responseWriter := &test.MockResponseWriter{
WriteMsgFunc: func(m *dns.Msg) error { return nil },
}
failed := false
resolver.deactivate = func(error) {
failed = true
// After deactivation, make the mock client work again
mockClient.err = nil
}
reactivated := false
@@ -154,7 +151,7 @@ func TestUpstreamResolver_DeactivationReactivation(t *testing.T) {
reactivated = true
}
resolver.ServeDNS(responseWriter, new(dns.Msg).SetQuestion("one.one.one.one.", dns.TypeA))
resolver.ProbeAvailability()
if !failed {
t.Errorf("expected that resolving was deactivated")
@@ -173,11 +170,6 @@ func TestUpstreamResolver_DeactivationReactivation(t *testing.T) {
return
}
if resolver.failsCount.Load() != 0 {
t.Errorf("fails count after reactivation should be 0")
return
}
if resolver.disabled {
t.Errorf("should be enabled")
}

Some files were not shown because too many files have changed in this diff Show More