Compare commits

...
Author SHA1 Message Date
Nate Brown 357b129749 Bind sshd, dns and stats listeners in the reload so a bind failure fails it 2026-10-19 09:57:21 -05:00
Matt RichardsonandClaude Fable 5.1 8a4bd8b10f Report ssh command failures to the surface (#1959)
smoke-extra / freebsd-amd64 (push) Failing after 14s
smoke-extra / linux-amd64-ipv6disable (push) Failing after 13s
smoke-extra / netbsd-amd64 (push) Failing after 14s
smoke-extra / openbsd-amd64 (push) Failing after 15s
smoke-extra / linux-386 (push) Failing after 14s
smoke / Run multi node smoke test (push) Failing after 1m46s
Build and test / Static checks (push) Successful in 2m38s
Build and test / Test linux (push) Failing after 1m12s
Build and test / Test linux-pkcs11 (push) Failing after 2m6s
Build and test / Test linux-boringcrypto (push) Failing after 3m0s
Build and test / Test linux-fips140 (push) Failing after 2m56s
Build and test / Cross-build linux-arm (push) Successful in 3m10s
Build and test / Cross-build linux-mips (push) Successful in 3m55s
Build and test / Cross-build linux-other (push) Successful in 3m15s
Build and test / Cross-build windows (push) Successful in 1m5s
Build and test / Cross-build freebsd (push) Successful in 1m37s
Build and test / Cross-build netbsd (push) Successful in 1m36s
Build and test / Cross-build openbsd (push) Successful in 1m35s
Build and test / Cross-build mobile (push) Successful in 3m32s
smoke-extra / Run windows smoke test (push) Canceled after 0s
smoke / Run self traffic smoke test on macOS (push) Canceled after 0s
Build and test / Test macos (push) Canceled after 0s
Build and test / Test windows (push) Canceled after 0s
Build and test / CI status (push) Canceled after 0s
Report ssh command failures to the surface (#1959)

The ssh debug commands told the user when something went wrong but
never told the surface that ran them, so `ssh host cmd` always exited 0
and a few failures were not reported at all. A failed command now
returns its error, a failure the user has already heard about is a
commands.Failure, and the ssh exec path exits 1 for either. The command
bodies also lose their ssh prefixes and the duplication they carried
over from ssh.go; both goldens pin that nothing else changed.

- Dispatch returns the callback's error, or shlex's; the ssh session
  logs it at debug, and not at all for a Failure the user already saw
- commands.Fail writes a message and returns it as a Failure in one step;
  an unknown command, `help` for a command that is not there, a failed
  reload and an unbalanced quote are failures
- ssh exec exit-status is 1 for a command, flag or failure error, 0
  otherwise; the e2e golden records the exit status for every case
- Clone rebinds help to the copy, so an ssh session's help lists logout
  and `help logout` and `logout -h` answer; sshd.LogoutCommand is the
  session's logout, exported for tests
- "did not understand" names the first word only, the one looked up;
  DispatchArgs takes just the tokens
- log-format writes its refusal to the user, as log-level does
- print-relays emits the relay error as a string and omits the key when
  there is none, where it printed {} or "Error":null
- list-hostmap and list-lighthouse-addrmap -json return the encode
  error; print-cert reports a certificate that will not marshal
- a flags type that is not the one registered is an internal error, not
  a silent success
- command functions and flag types are <name>Command and <name>Flags;
  parseVpnAddr and listHostMapFlagSet replace the repeated parse block
  and flag set; close-tunnel sends through sendCloseTunnel

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
2026-10-19 10:48:26 -04:00
Matt RichardsonandClaude Fable 5.1 b426843f88 Extract the ssh debug commands into a transport neutral commands package (#1958)
Migrate commands from ssh into a transport neutral package

 - add golden tests validating that output remains unchanged
 - move commands to commands.go

Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
2026-10-19 10:25:28 -04:00
Wade Simmons cda3710d5a fix error return in handleOutsideMessagePacket (#1970)
smoke-extra / freebsd-amd64 (push) Failing after 17s
smoke-extra / linux-amd64-ipv6disable (push) Failing after 16s
smoke-extra / netbsd-amd64 (push) Failing after 16s
smoke-extra / openbsd-amd64 (push) Failing after 16s
smoke-extra / linux-386 (push) Failing after 17s
smoke / Run multi node smoke test (push) Failing after 1m39s
Build and test / Static checks (push) Successful in 45s
Build and test / Test linux (push) Failing after 1m18s
Build and test / Test linux-pkcs11 (push) Failing after 2m5s
Build and test / Test linux-boringcrypto (push) Failing after 3m1s
Build and test / Test linux-fips140 (push) Failing after 2m55s
Build and test / Cross-build linux-arm (push) Successful in 3m0s
Build and test / Cross-build linux-mips (push) Successful in 3m42s
Build and test / Cross-build linux-other (push) Successful in 3m5s
Build and test / Cross-build windows (push) Successful in 1m2s
Build and test / Cross-build freebsd (push) Successful in 1m32s
Build and test / Cross-build netbsd (push) Successful in 1m31s
Build and test / Cross-build openbsd (push) Successful in 1m32s
Build and test / Cross-build mobile (push) Successful in 3m21s
smoke-extra / Run windows smoke test (push) Canceled after 0s
smoke / Run self traffic smoke test on macOS (push) Canceled after 0s
Build and test / Test macos (push) Canceled after 0s
Build and test / Test windows (push) Canceled after 0s
Build and test / CI status (push) Canceled after 0s
Introduced by #1961, we should still return early for invalid packets
2026-10-17 11:09:25 -04:00
John Maguire c26871fe08 service: Close listeners and tear down the netstack on shutdown (#1954) 2026-10-17 11:08:29 -04:00
Caleb Jasik 7daacc6a48 overlay/darwin: test tun framing and zero-alloc Read/Write (#1906)
smoke-extra / freebsd-amd64 (push) Failing after 16s
smoke-extra / linux-amd64-ipv6disable (push) Failing after 16s
smoke-extra / netbsd-amd64 (push) Failing after 17s
smoke-extra / openbsd-amd64 (push) Failing after 17s
smoke-extra / linux-386 (push) Failing after 16s
smoke / Run multi node smoke test (push) Failing after 1m41s
Build and test / Static checks (push) Successful in 45s
Build and test / Test linux (push) Failing after 1m47s
Build and test / Test linux-pkcs11 (push) Failing after 2m6s
Build and test / Test linux-boringcrypto (push) Failing after 2m57s
Build and test / Test linux-fips140 (push) Failing after 2m53s
Build and test / Cross-build linux-arm (push) Successful in 3m7s
Build and test / Cross-build linux-mips (push) Successful in 3m55s
Build and test / Cross-build linux-other (push) Successful in 3m13s
Build and test / Cross-build windows (push) Successful in 1m4s
Build and test / Cross-build freebsd (push) Successful in 1m35s
Build and test / Cross-build netbsd (push) Successful in 1m35s
Build and test / Cross-build openbsd (push) Successful in 1m34s
Build and test / Cross-build mobile (push) Successful in 3m26s
smoke-extra / Run windows smoke test (push) Canceled after 0s
smoke / Run self traffic smoke test on macOS (push) Canceled after 0s
Build and test / Test macos (push) Canceled after 0s
Build and test / Test windows (push) Canceled after 0s
Build and test / CI status (push) Canceled after 0s
2026-10-16 18:12:58 -05:00
John Maguire 428dad94b3 Reject v2 cert names and groups at signing the parser rejects (#1967)
smoke-extra / freebsd-amd64 (push) Failing after 18s
smoke-extra / linux-amd64-ipv6disable (push) Failing after 18s
smoke-extra / netbsd-amd64 (push) Failing after 16s
smoke-extra / openbsd-amd64 (push) Failing after 17s
smoke-extra / linux-386 (push) Failing after 17s
smoke / Run multi node smoke test (push) Failing after 1m41s
Build and test / Static checks (push) Successful in 45s
Build and test / Test linux (push) Failing after 1m19s
Build and test / Test linux-pkcs11 (push) Failing after 2m5s
Build and test / Test linux-boringcrypto (push) Failing after 2m56s
Build and test / Test linux-fips140 (push) Failing after 2m51s
Build and test / Cross-build linux-arm (push) Successful in 3m0s
Build and test / Cross-build linux-mips (push) Successful in 3m43s
Build and test / Cross-build linux-other (push) Successful in 3m7s
Build and test / Cross-build windows (push) Successful in 1m2s
Build and test / Cross-build freebsd (push) Successful in 1m32s
Build and test / Cross-build netbsd (push) Successful in 1m33s
Build and test / Cross-build openbsd (push) Successful in 1m32s
Build and test / Cross-build mobile (push) Successful in 3m23s
smoke-extra / Run windows smoke test (push) Canceled after 0s
smoke / Run self traffic smoke test on macOS (push) Canceled after 0s
Build and test / Test macos (push) Canceled after 0s
Build and test / Test windows (push) Canceled after 0s
Build and test / CI status (push) Canceled after 0s
2026-10-16 18:05:51 -04:00
John Maguire 5bf666dd90 Treat uppercase blocked fingerprints the same as lowercase (#1964)
- Error for invalid on startup 
- Still block the valid blocklist entries when a reload has a bad one
2026-10-16 16:28:37 -04:00
John Maguire 49207da077 Remove the ssh debug profiling commands (#1965)
We plan to move this functionality into the debug socket.
2026-10-16 15:47:56 -04:00
Jack Doan bbc49669b7 fix noisy log line (#1961) 2026-10-16 12:08:55 -05:00
Nate Brown 915495d5f4 Fail the reload on a startup only value that would fail a restart, and stop asking for tun.dev restarts that change nothing (#1953)
smoke-extra / freebsd-amd64 (push) Failing after 25s
smoke-extra / linux-amd64-ipv6disable (push) Failing after 13s
smoke-extra / netbsd-amd64 (push) Failing after 11s
smoke-extra / openbsd-amd64 (push) Failing after 12s
smoke-extra / linux-386 (push) Failing after 14s
smoke / Run multi node smoke test (push) Failing after 1m43s
Build and test / Static checks (push) Successful in 2m25s
Build and test / Test linux (push) Failing after 1m28s
Build and test / Test linux-pkcs11 (push) Failing after 2m5s
Build and test / Test linux-boringcrypto (push) Failing after 2m56s
Build and test / Test linux-fips140 (push) Failing after 3m3s
Build and test / Cross-build linux-arm (push) Successful in 3m14s
Build and test / Cross-build linux-mips (push) Successful in 3m48s
Build and test / Cross-build linux-other (push) Successful in 3m13s
Build and test / Cross-build windows (push) Successful in 1m3s
Build and test / Cross-build freebsd (push) Successful in 1m34s
Build and test / Cross-build netbsd (push) Successful in 1m36s
Build and test / Cross-build openbsd (push) Successful in 1m37s
Build and test / Cross-build mobile (push) Successful in 3m33s
smoke / Run self traffic smoke test on macOS (push) Canceled after 0s
Build and test / Test macos (push) Canceled after 0s
Build and test / Test windows (push) Canceled after 0s
Build and test / CI status (push) Canceled after 0s
smoke-extra / Run windows smoke test (push) Canceled after 0s
2026-10-15 23:34:09 -05:00
Nate Brown a203d044bb Fail the reload when a config entry is skipped with a warning (#1952) 2026-10-15 23:22:47 -05:00
Nate Brown 316bf87780 Don't restart sshd when only sshd.sandbox_dir changes (#1948) 2026-10-15 22:12:12 -05:00
Nate Brown 1a0ac25db0 Report how a config reload went: failures, restarts required, and the keys that need one (#1935) 2026-10-15 15:36:39 -05:00
John Maguire 42597dda7d service: Don't hold listener lock while handing off a connection (#1939) 2026-10-15 15:14:39 -04:00
John Maguire afbdc87ed9 Update gvisor and bump to go1.27 (#1940)
gvisor's arm64 race-build compare-and-swap (pkg/sync/race_arm64.s)
loaded the new value into the register meant for the old one (R1
instead of R2), so it never succeeded. Sleepers could not commit to
parking and spun: under -race on arm64 every idle service netstack
busied about 14 cores, and TestService's static host lookup of
localhost missed its 250ms timeout and stalled for the 30s static_map
cadence. Fixed in gvisor 1b2946c4fa.

Go 1.27.2 is now the minimum because gvisor's go branch requires it.
golangci-lint is bumped to v2.14, the first release that can read
go1.27.2 export data.
2026-10-14 21:30:59 -04:00
Nate Brown 94372f4f51 Don't delete routes we never installed on darwin and the BSDs (#1933) 2026-10-13 16:34:38 -05:00
John Maguire 42280f04ce darwin: Assign host address to the utun and route vpn explicitly (#1931)
Tailscale with an exit node treats every interface address prefix as an
attached LAN and either claims it or excludes it, and configd then rewrites
the matching route away from nebula's utun. Interfaces holding a single
address are skipped, so assign /32 and /128 and add the network route
ourselves for both families.
2026-10-13 13:50:37 -04:00
Nate Brown 11c1917387 Warn when an unsafe_route is already routed through another interface on darwin (#1925) 2026-10-12 16:57:14 -05:00
Nate Brown 5ba7aff4b3 ssh: Only restart sshd on reload when its config or host key changed (#1923) 2026-10-12 16:56:51 -05:00
Ryan Huber 8a2de9b1b5 examples: set socket buffers and raise tun tx_queue (#1928)
* examples: set socket buffers and raise tun tx_queue

* examples: use the default tcp max window for socket buffers

* examples: note buffers are linux only, clarify tx_queue default

* tun: default tx_queue to 5000
2026-10-11 00:30:07 -04:00
Nate Brown a8c52f5c13 overlay: Handle routes Windows already has, and parse quoted route numbers (#1924) 2026-10-08 18:52:53 -05:00
Jack Doan 13af8dd6f3 Cap configured MTUs at 8937 so relayed sends never outgrow their buffers (#1899) 2026-10-08 15:41:08 -05:00
Nate Brown 5bb0622192 stats: Follow stats.interval on reload, interface gauges ride the capture loop (#1919) 2026-10-08 10:00:48 -05:00
Nate Brown f3705cbb67 graphite: Time out a send, and abandon it on stop (#1920) 2026-10-07 18:01:09 -05:00
Nate Brown 94be454f4f lighthouse: Recompute relays for me when relay.am_relay changes (#1917) 2026-10-07 14:06:27 -05:00
Nate Brown 5f26c1a7e5 Add a simple rate limiter with a helper to use it in logging (#1915) 2026-10-07 10:53:37 -05:00
Nate Brown d47c8c71f1 lighthouse: Send an update when the relays we advertise change on reload (#1916) 2026-10-07 10:52:42 -05:00
Nate Brown f57539a163 overlay/linux: Keep unsafe routes when a reload only changes tun.mtu (#1918) 2026-10-07 10:43:11 -05:00
John Maguire 8d8ae0f744 Disable DNS registration on the Windows tunnel adapter (#1914)
On a domain-joined Windows host the DNS Client registers every adapter's
addresses under the host name in AD DNS when "Register this connection's
addresses in DNS" is on, which is the default for a new adapter. When the
Nebula adapter registers, LAN clients that are not on the overlay resolve
the host name to its overlay IP and RDP, SMB and PowerShell Remoting to it
fail intermittently, with nothing pointing at the tunnel as the cause.
Windows registers on address assignment, periodically, and on every
ipconfig /registerdns.

Right after the wintun adapter is created and before Activate assigns any
address, write RegistrationEnabled=0, DisableDynamicUpdate=1 and
MaxNumberOfAddressesToRegister=0 under both the Tcpip and Tcpip6
Parameters\Interfaces\{GUID} keys. The adapter checkbox alone has proven
insufficient in practice, as has disabling dynamic updates alone, so all
three are set, matching what Tailscale and NetBird do. A failure to write
the keys is logged as a warning and does not fail adapter creation.
2026-10-07 11:42:08 -04:00
Matt Richardson f9b0853476 Attach a pending handshake's RemoteList when the handshake starts (#1909)
StartHandshake returned a hostinfo with no RemoteList; handleOutbound attached
one lazily on its first tick. Anything that touched the pending host before
then, notably create-tunnel -address, dereferenced nil and took the daemon
down. StartHandshake now attaches the lighthouse's list as it builds the
hostinfo. To keep the surrounding locking sound, a lighthouse now resolves a
punch target before taking the querier's RemoteList lock, and closeTunnel
deletes the host and evicts its RemoteList under a single hostmap write lock so
a completing handshake cannot land in between. The lock order, now documented
on the handshake manager mutex, is hostmap, then handshake manager, then
lighthouse, then RemoteList; no RemoteList lock is ever held across a hostmap
lookup.
2026-10-07 11:22:50 -04:00
Nate Brown d5062dacf0 lighthouse: Never block in QueryServer, drop queries when the channel is full (#1910) 2026-09-30 15:26:34 -05:00
110 changed files with 7276 additions and 1649 deletions

No files matched your search

+2
View File
@@ -0,0 +1,2 @@
# Goldens are compared byte for byte, so they must come out of git the way they went in.
*.golden -text
+3 -3
View File
@@ -14,7 +14,7 @@ jobs:
- uses: actions/setup-go@v7
with:
go-version: '1.26'
go-version: '1.27'
check-latest: true
- name: Build
@@ -40,7 +40,7 @@ jobs:
- uses: actions/setup-go@v7
with:
go-version: '1.26'
go-version: '1.27'
check-latest: true
- name: Build
@@ -83,7 +83,7 @@ jobs:
- uses: actions/setup-go@v7
with:
go-version: '1.26'
go-version: '1.27'
check-latest: true
# GitHub holds ARNs, not credentials, and ARNs outlive a rotation
+3 -3
View File
@@ -34,7 +34,7 @@ jobs:
- uses: actions/setup-go@v7
with:
go-version: '1.26'
go-version: '1.27'
check-latest: true
- name: add hashicorp source
@@ -66,7 +66,7 @@ jobs:
- uses: actions/setup-go@v7
with:
go-version: '1.26'
go-version: '1.27'
check-latest: true
- name: add hashicorp source
@@ -92,7 +92,7 @@ jobs:
- uses: actions/setup-go@v7
with:
go-version: '1.26'
go-version: '1.27'
check-latest: true
# WSL2 + Ubuntu so the smoke can run a real linux peer with its own
+2 -2
View File
@@ -22,7 +22,7 @@ jobs:
- uses: actions/setup-go@v7
with:
go-version: '1.26'
go-version: '1.27'
check-latest: true
- name: Smoke Docker
@@ -51,7 +51,7 @@ jobs:
- uses: actions/setup-go@v7
with:
go-version: '1.26'
go-version: '1.27'
check-latest: true
- name: build
+4 -4
View File
@@ -22,7 +22,7 @@ jobs:
- uses: actions/setup-go@v7
with:
go-version: '1.26'
go-version: '1.27'
check-latest: true
- name: Install goimports
@@ -42,7 +42,7 @@ jobs:
- name: golangci-lint
uses: golangci/golangci-lint-action@v9
with:
version: v2.12
version: v2.14
test:
name: Test ${{ matrix.name }}
@@ -87,7 +87,7 @@ jobs:
- uses: actions/setup-go@v7
with:
go-version: '1.26'
go-version: '1.27'
check-latest: true
- name: Build
@@ -132,7 +132,7 @@ jobs:
- uses: actions/setup-go@v7
with:
go-version: '1.26'
go-version: '1.27'
check-latest: true
- name: Build ${{ matrix.name }}
+14
View File
@@ -7,6 +7,20 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
## [Unreleased]
### Removed
- The SSH debug server's `start-cpu-profile`, `stop-cpu-profile`, `save-heap-profile`, `save-mutex-profile`, and
`mutex-profile-fraction` commands, and `sshd.sandbox_dir`, which is now ignored if set.
### Fixed
- Signing a v2 certificate with an empty name, a name longer than 253 bytes, or an empty group now fails.
These certificates were signed successfully but could not be loaded.
- `create-tunnel -address <ip:port>` panicked on a nil RemoteList and, over the ssh debug server, took nebula
down with it. A hostinfo returned by StartHandshake had no RemoteList until the handshake loop's first
pass attached one, and `create-tunnel -address` set the remote before that. The list is now attached when
the handshake starts, so a pending hostinfo never exists without one.
## [1.11.1] - 2026-08-21
See the [v1.11.1](https://github.com/slackhq/nebula/milestone/30?closed=1) milestone for a complete list of changes.
+10 -1
View File
@@ -389,7 +389,16 @@ func (c *certificateV2) fromTBSCertificate(t *TBSCertificate) error {
}
func (c *certificateV2) validate() error {
// Empty names are allowed
// Mirror the name and group rules in unmarshalDetails so signing fails instead of producing an unparseable certificate
if c.details.name == "" || len(c.details.name) > MaxNameLength {
return NewErrInvalidCertificateProperties("name must be between 1 and %d bytes", MaxNameLength)
}
for _, group := range c.details.groups {
if group == "" {
return NewErrInvalidCertificateProperties("groups may not be empty")
}
}
if len(c.publicKey) == 0 {
return ErrInvalidPublicKey
+45
View File
@@ -6,6 +6,7 @@ import (
"crypto/elliptic"
"crypto/rand"
"net/netip"
"strings"
"testing"
"time"
@@ -135,3 +136,47 @@ func TestCertificate_SignP256_AlwaysNormalized(t *testing.T) {
assert.True(t, normie)
}
}
func TestCertificateV2_SignRejectsUnparseable(t *testing.T) {
_, priv, err := ed25519.GenerateKey(rand.Reader)
require.NoError(t, err)
var invalid *ErrInvalidCertificateProperties
newTBS := func() TBSCertificate {
return TBSCertificate{
Version: Version2,
Name: "testing",
Networks: []netip.Prefix{mustParsePrefixUnmapped("10.1.1.1/24")},
Groups: []string{"test-group1"},
NotBefore: time.Now().Add(-time.Minute),
NotAfter: time.Now().Add(time.Minute),
PublicKey: []byte("1234567890abcedfghij1234567890ab"),
Curve: Curve_CURVE25519,
IsCA: true,
}
}
tbs := newTBS()
tbs.Name = strings.Repeat("a", MaxNameLength)
c, err := tbs.Sign(nil, Curve_CURVE25519, priv)
require.NoError(t, err)
b, err := c.Marshal()
require.NoError(t, err)
_, err = unmarshalCertificateV2(b, nil, Curve_CURVE25519)
require.NoError(t, err)
tbs = newTBS()
tbs.Name = strings.Repeat("a", MaxNameLength+1)
_, err = tbs.Sign(nil, Curve_CURVE25519, priv)
require.ErrorAs(t, err, &invalid)
tbs = newTBS()
tbs.Name = ""
_, err = tbs.Sign(nil, Curve_CURVE25519, priv)
require.ErrorAs(t, err, &invalid)
tbs = newTBS()
tbs.Groups = []string{"test-group1", ""}
_, err = tbs.Sign(nil, Curve_CURVE25519, priv)
require.ErrorAs(t, err, &invalid)
}
+2 -1
View File
@@ -85,9 +85,10 @@ func main() {
fmt.Printf("failed to apply logging config: %s", err)
os.Exit(1)
}
c.RegisterReloadCallback(func(c *config.C) {
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
if err := logging.ApplyConfig(l, c); err != nil {
l.Error("Failed to reconfigure logger on reload", "error", err)
r.Failed()
}
})
+2 -1
View File
@@ -34,9 +34,10 @@ func (p *program) Start(s service.Service) error {
if err := logging.ApplyConfig(l, c); err != nil {
return fmt.Errorf("failed to apply logging config: %s", err)
}
c.RegisterReloadCallback(func(c *config.C) {
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
if err := logging.ApplyConfig(l, c); err != nil {
l.Error("Failed to reconfigure logger on reload", "error", err)
r.Failed()
}
})
+2 -1
View File
@@ -71,9 +71,10 @@ func main() {
fmt.Printf("failed to apply logging config: %s", err)
os.Exit(1)
}
c.RegisterReloadCallback(func(c *config.C) {
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
if err := logging.ApplyConfig(l, c); err != nil {
l.Error("Failed to reconfigure logger on reload", "error", err)
r.Failed()
}
})
+710
View File
@@ -0,0 +1,710 @@
package nebula
import (
"bytes"
"encoding/json"
"flag"
"fmt"
"log/slog"
"maps"
"net/netip"
"sort"
"strings"
"github.com/slackhq/nebula/commands"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/logging"
)
// The commands nebula serves. attachCommands builds the one registry every surface
// reads. See the commands package for what a Command is.
type listHostMapFlags struct {
Json bool
Pretty bool
ByIndex bool
}
// listHostMapFlagSet is the flag set list-hostmap and list-pending-hostmap share.
func listHostMapFlagSet() (*flag.FlagSet, any) {
fl := flag.NewFlagSet("", flag.ContinueOnError)
s := listHostMapFlags{}
fl.BoolVar(&s.Json, "json", false, "outputs as json with more information")
fl.BoolVar(&s.Pretty, "pretty", false, "pretty prints json, assumes -json")
fl.BoolVar(&s.ByIndex, "by-index", false, "gets all hosts in the hostmap from the index table")
return fl, &s
}
type printCertFlags struct {
Json bool
Pretty bool
Raw bool
}
type printTunnelFlags struct {
Pretty bool
}
type changeRemoteFlags struct {
Address string
}
type closeTunnelFlags struct {
LocalOnly bool
}
type createTunnelFlags struct {
Address string
}
type deviceInfoFlags struct {
Json bool
Pretty bool
}
func attachCommands(l *slog.Logger, c *config.C, reg *commands.Registry, f *Interface) {
reg.Register(&commands.Command{
Name: commands.ListHostmap,
ShortDescription: "List all known previously connected hosts",
Flags: listHostMapFlagSet,
Callback: func(fs any, a []string, w commands.StringWriter) error {
return listHostMapCommand(f.hostMap, fs, w)
},
})
reg.Register(&commands.Command{
Name: commands.ListPendingHostmap,
ShortDescription: "List all handshaking hosts",
Flags: listHostMapFlagSet,
Callback: func(fs any, a []string, w commands.StringWriter) error {
return listHostMapCommand(f.handshakeManager, fs, w)
},
})
reg.Register(&commands.Command{
Name: commands.ListLighthouseAddrmap,
ShortDescription: "List all lighthouse map entries",
Flags: func() (*flag.FlagSet, any) {
fl := flag.NewFlagSet("", flag.ContinueOnError)
s := listHostMapFlags{}
fl.BoolVar(&s.Json, "json", false, "outputs as json with more information")
fl.BoolVar(&s.Pretty, "pretty", false, "pretty prints json, assumes -json")
return fl, &s
},
Callback: func(fs any, a []string, w commands.StringWriter) error {
return listLighthouseAddrmapCommand(f.lightHouse, fs, w)
},
})
reg.Register(&commands.Command{
Name: commands.Reload,
ShortDescription: "Reloads configuration from disk, same as sending HUP to the process",
Callback: func(fs any, a []string, w commands.StringWriter) error {
return reloadCommand(l, c, w)
},
})
reg.Register(&commands.Command{
Name: commands.LogLevel,
ShortDescription: "Gets or sets the current log level",
Callback: func(fs any, a []string, w commands.StringWriter) error {
return logLevelCommand(l, fs, a, w)
},
})
reg.Register(&commands.Command{
Name: commands.LogFormat,
ShortDescription: "Gets or sets the current log format",
Callback: func(fs any, a []string, w commands.StringWriter) error {
return logFormatCommand(l, fs, a, w)
},
})
reg.Register(&commands.Command{
Name: commands.Version,
ShortDescription: "Prints the currently running version of nebula",
Callback: func(fs any, a []string, w commands.StringWriter) error {
return versionCommand(f, fs, a, w)
},
})
reg.Register(&commands.Command{
Name: commands.DeviceInfo,
ShortDescription: "Prints information about the network device.",
Flags: func() (*flag.FlagSet, any) {
fl := flag.NewFlagSet("", flag.ContinueOnError)
s := deviceInfoFlags{}
fl.BoolVar(&s.Json, "json", false, "outputs as json with more information")
fl.BoolVar(&s.Pretty, "pretty", false, "pretty prints json, assumes -json")
return fl, &s
},
Callback: func(fs any, a []string, w commands.StringWriter) error {
return deviceInfoCommand(f, fs, w)
},
})
reg.Register(&commands.Command{
Name: commands.PrintCert,
ShortDescription: "Prints the current certificate being used or the certificate for the provided vpn addr",
Flags: func() (*flag.FlagSet, any) {
fl := flag.NewFlagSet("", flag.ContinueOnError)
s := printCertFlags{}
fl.BoolVar(&s.Json, "json", false, "outputs as json")
fl.BoolVar(&s.Pretty, "pretty", false, "pretty prints json, assumes -json")
fl.BoolVar(&s.Raw, "raw", false, "raw prints the PEM encoded certificate, not compatible with -json or -pretty")
return fl, &s
},
Callback: func(fs any, a []string, w commands.StringWriter) error {
return printCertCommand(f, fs, a, w)
},
})
reg.Register(&commands.Command{
Name: commands.PrintTunnel,
ShortDescription: "Prints json details about a tunnel for the provided vpn addr",
Flags: func() (*flag.FlagSet, any) {
fl := flag.NewFlagSet("", flag.ContinueOnError)
s := printTunnelFlags{}
fl.BoolVar(&s.Pretty, "pretty", false, "pretty prints json")
return fl, &s
},
Callback: func(fs any, a []string, w commands.StringWriter) error {
return printTunnelCommand(f, fs, a, w)
},
})
reg.Register(&commands.Command{
Name: commands.PrintRelays,
ShortDescription: "Prints json details about all relay info",
Flags: func() (*flag.FlagSet, any) {
fl := flag.NewFlagSet("", flag.ContinueOnError)
s := printTunnelFlags{}
fl.BoolVar(&s.Pretty, "pretty", false, "pretty prints json")
return fl, &s
},
Callback: func(fs any, a []string, w commands.StringWriter) error {
return printRelaysCommand(f, fs, a, w)
},
})
reg.Register(&commands.Command{
Name: commands.ChangeRemote,
ShortDescription: "Changes the remote address used in the tunnel for the provided vpn addr",
Flags: func() (*flag.FlagSet, any) {
fl := flag.NewFlagSet("", flag.ContinueOnError)
s := changeRemoteFlags{}
fl.StringVar(&s.Address, "address", "", "The new remote address, ip:port")
return fl, &s
},
Callback: func(fs any, a []string, w commands.StringWriter) error {
return changeRemoteCommand(f, fs, a, w)
},
})
reg.Register(&commands.Command{
Name: commands.CloseTunnel,
ShortDescription: "Closes a tunnel for the provided vpn addr",
Flags: func() (*flag.FlagSet, any) {
fl := flag.NewFlagSet("", flag.ContinueOnError)
s := closeTunnelFlags{}
fl.BoolVar(&s.LocalOnly, "local-only", false, "Disables notifying the remote that the tunnel is shutting down")
return fl, &s
},
Callback: func(fs any, a []string, w commands.StringWriter) error {
return closeTunnelCommand(f, fs, a, w)
},
})
reg.Register(&commands.Command{
Name: commands.CreateTunnel,
ShortDescription: "Creates a tunnel for the provided vpn address",
Help: "The lighthouses will be queried for real addresses but you can provide one as well.",
Flags: func() (*flag.FlagSet, any) {
fl := flag.NewFlagSet("", flag.ContinueOnError)
s := createTunnelFlags{}
fl.StringVar(&s.Address, "address", "", "Optionally provide a real remote address, ip:port ")
return fl, &s
},
Callback: func(fs any, a []string, w commands.StringWriter) error {
return createTunnelCommand(f, fs, a, w)
},
})
reg.Register(&commands.Command{
Name: commands.QueryLighthouse,
ShortDescription: "Query the lighthouses for the provided vpn address",
Help: "This command is asynchronous. Only currently known udp addresses will be printed.",
Callback: func(fs any, a []string, w commands.StringWriter) error {
return queryLighthouseCommand(f, fs, a, w)
},
})
}
func listHostMapCommand(hl controlHostLister, a any, w commands.StringWriter) error {
fs, ok := a.(*listHostMapFlags)
if !ok {
return fmt.Errorf("internal error: expected *listHostMapFlags but was %T", a)
}
var hm []ControlHostInfo
if fs.ByIndex {
hm = listHostMapIndexes(hl)
} else {
hm = listHostMapHosts(hl)
}
sort.Slice(hm, func(i, j int) bool {
return hm[i].VpnAddrs[0].Compare(hm[j].VpnAddrs[0]) < 0
})
if fs.Json || fs.Pretty {
js := json.NewEncoder(w.GetWriter())
if fs.Pretty {
js.SetIndent("", " ")
}
if err := js.Encode(hm); err != nil {
return err
}
} else {
for _, v := range hm {
err := w.WriteLine(fmt.Sprintf("%s: %s", v.VpnAddrs, v.RemoteAddrs))
if err != nil {
return err
}
}
}
return nil
}
func listLighthouseAddrmapCommand(lightHouse *LightHouse, a any, w commands.StringWriter) error {
fs, ok := a.(*listHostMapFlags)
if !ok {
return fmt.Errorf("internal error: expected *listHostMapFlags but was %T", a)
}
type lighthouseInfo struct {
VpnAddr string `json:"vpnAddr"`
Addrs *CacheMap `json:"addrs"`
}
lightHouse.RLock()
addrMap := make([]lighthouseInfo, len(lightHouse.addrMap))
x := 0
for k, v := range lightHouse.addrMap {
addrMap[x] = lighthouseInfo{
VpnAddr: k.String(),
Addrs: v.CopyCache(),
}
x++
}
lightHouse.RUnlock()
sort.Slice(addrMap, func(i, j int) bool {
return strings.Compare(addrMap[i].VpnAddr, addrMap[j].VpnAddr) < 0
})
if fs.Json || fs.Pretty {
js := json.NewEncoder(w.GetWriter())
if fs.Pretty {
js.SetIndent("", " ")
}
if err := js.Encode(addrMap); err != nil {
return err
}
} else {
for _, v := range addrMap {
b, err := json.Marshal(v.Addrs)
if err != nil {
return err
}
err = w.WriteLine(fmt.Sprintf("%s: %s", v.VpnAddr, string(b)))
if err != nil {
return err
}
}
}
return nil
}
// parseVpnAddr parses the vpn address a command takes as its first operand, failing to w
// when there is none or it does not parse. noun is how the command's messages name it,
// "vpn address" or "vpn addr", which the output golden pins per command.
func parseVpnAddr(a []string, noun string, w commands.StringWriter) (netip.Addr, error) {
if len(a) == 0 {
return netip.Addr{}, commands.Fail(w, "No vpn address was provided")
}
vpnAddr, err := netip.ParseAddr(a[0])
if err != nil {
return netip.Addr{}, commands.Fail(w, fmt.Sprintf("The provided %s could not be parsed: %s", noun, a[0]))
}
return vpnAddr, nil
}
func versionCommand(ifce *Interface, fs any, a []string, w commands.StringWriter) error {
return w.WriteLine(ifce.version)
}
func queryLighthouseCommand(ifce *Interface, fs any, a []string, w commands.StringWriter) error {
vpnAddr, err := parseVpnAddr(a, "vpn address", w)
if err != nil {
return err
}
var cm *CacheMap
rl := ifce.lightHouse.Query(vpnAddr)
if rl != nil {
cm = rl.CopyCache()
}
return json.NewEncoder(w.GetWriter()).Encode(cm)
}
func closeTunnelCommand(ifce *Interface, fs any, a []string, w commands.StringWriter) error {
flags, ok := fs.(*closeTunnelFlags)
if !ok {
return fmt.Errorf("internal error: expected *closeTunnelFlags but was %T", fs)
}
vpnAddr, err := parseVpnAddr(a, "vpn address", w)
if err != nil {
return err
}
hostInfo := ifce.hostMap.QueryVpnAddr(vpnAddr)
if hostInfo == nil {
return commands.Fail(w, fmt.Sprintf("Could not find tunnel for vpn address: %v", a[0]))
}
if !flags.LocalOnly {
ifce.sendCloseTunnel(hostInfo)
}
ifce.closeTunnel(hostInfo)
return w.WriteLine("Closed")
}
func createTunnelCommand(ifce *Interface, fs any, a []string, w commands.StringWriter) error {
flags, ok := fs.(*createTunnelFlags)
if !ok {
return fmt.Errorf("internal error: expected *createTunnelFlags but was %T", fs)
}
vpnAddr, err := parseVpnAddr(a, "vpn address", w)
if err != nil {
return err
}
hostInfo := ifce.hostMap.QueryVpnAddr(vpnAddr)
if hostInfo != nil {
return commands.Fail(w, "Tunnel already exists")
}
hostInfo = ifce.handshakeManager.QueryVpnAddr(vpnAddr)
if hostInfo != nil {
return commands.Fail(w, "Tunnel already handshaking")
}
var addr netip.AddrPort
if flags.Address != "" {
addr, err = netip.ParseAddrPort(flags.Address)
if err != nil {
return commands.Fail(w, "Address could not be parsed")
}
}
// Set the remote from the callback, which StartHandshake runs under the handshake manager lock
// before the hostinfo reaches the handshake loop or the static-host trigger, so the address is
// in the RemoteList before the first handshake is sent.
ifce.handshakeManager.StartHandshake(vpnAddr, func(hh *HandshakeHostInfo) {
if addr.IsValid() {
hh.hostinfo.SetRemote(addr)
}
})
return w.WriteLine("Created")
}
func changeRemoteCommand(ifce *Interface, fs any, a []string, w commands.StringWriter) error {
flags, ok := fs.(*changeRemoteFlags)
if !ok {
return fmt.Errorf("internal error: expected *changeRemoteFlags but was %T", fs)
}
vpnAddr, err := parseVpnAddr(a, "vpn address", w)
if err != nil {
return err
}
if flags.Address == "" {
return commands.Fail(w, "No address was provided")
}
addr, err := netip.ParseAddrPort(flags.Address)
if err != nil {
return commands.Fail(w, "Address could not be parsed")
}
hostInfo := ifce.hostMap.QueryVpnAddr(vpnAddr)
if hostInfo == nil {
return commands.Fail(w, fmt.Sprintf("Could not find tunnel for vpn address: %v", a[0]))
}
hostInfo.SetRemote(addr)
return w.WriteLine("Changed")
}
func logLevelCommand(l *slog.Logger, fs any, a []string, w commands.StringWriter) error {
ctrl, ok := l.Handler().(interface {
GetLevel() slog.Level
SetLevel(slog.Level)
})
if !ok {
return commands.Fail(w, "Log level is not reconfigurable on this logger")
}
if len(a) == 0 {
return w.WriteLine(fmt.Sprintf("Log level is: %s", logging.LevelName(ctrl.GetLevel())))
}
level, err := logging.ParseLevel(strings.ToLower(a[0]))
if err != nil {
return commands.Fail(w, fmt.Sprintf("Unknown log level %s. Possible log levels: trace, debug, info, warn, error", a))
}
ctrl.SetLevel(level)
return w.WriteLine(fmt.Sprintf("Log level is: %s", logging.LevelName(ctrl.GetLevel())))
}
func logFormatCommand(l *slog.Logger, fs any, a []string, w commands.StringWriter) error {
ctrl, ok := l.Handler().(interface {
GetFormat() string
SetFormat(string) error
})
if !ok {
return commands.Fail(w, "Log format is not reconfigurable on this logger")
}
if len(a) == 0 {
return w.WriteLine(fmt.Sprintf("Log format is: %s", ctrl.GetFormat()))
}
if err := ctrl.SetFormat(strings.ToLower(a[0])); err != nil {
return commands.Fail(w, fmt.Sprintf("Unable to set log format: %s", err))
}
return w.WriteLine(fmt.Sprintf("Log format is: %s", ctrl.GetFormat()))
}
func printCertCommand(ifce *Interface, fs any, a []string, w commands.StringWriter) error {
args, ok := fs.(*printCertFlags)
if !ok {
return fmt.Errorf("internal error: expected *printCertFlags but was %T", fs)
}
cert := ifce.pki.getCertState().GetDefaultCertificate()
if len(a) > 0 {
vpnAddr, err := parseVpnAddr(a, "vpn addr", w)
if err != nil {
return err
}
hostInfo := ifce.hostMap.QueryVpnAddr(vpnAddr)
if hostInfo == nil {
return commands.Fail(w, fmt.Sprintf("Could not find tunnel for vpn addr: %v", a[0]))
}
cert = hostInfo.GetCert().Certificate
}
if args.Json || args.Pretty {
b, err := cert.MarshalJSON()
if err != nil {
return commands.Fail(w, fmt.Sprintf("Unable to marshal the certificate: %s", err))
}
if args.Pretty {
buf := new(bytes.Buffer)
if err := json.Indent(buf, b, "", " "); err != nil {
return commands.Fail(w, fmt.Sprintf("Unable to marshal the certificate: %s", err))
}
b = buf.Bytes()
}
return w.WriteBytes(b)
}
if args.Raw {
b, err := cert.MarshalPEM()
if err != nil {
return commands.Fail(w, fmt.Sprintf("Unable to marshal the certificate: %s", err))
}
return w.WriteBytes(b)
}
return w.WriteLine(cert.String())
}
func printRelaysCommand(ifce *Interface, fs any, a []string, w commands.StringWriter) error {
args, ok := fs.(*printTunnelFlags)
if !ok {
return fmt.Errorf("internal error: expected *printTunnelFlags but was %T", fs)
}
ifce.hostMap.RLock()
relays := maps.Clone(ifce.hostMap.Relays)
ifce.hostMap.RUnlock()
type RelayFor struct {
Error string `json:",omitempty"`
Type string
State string
PeerAddr netip.Addr
LocalIndex uint32
RemoteIndex uint32
RelayedThrough []netip.Addr
}
type RelayOutput struct {
NebulaAddr netip.Addr
RelayForAddrs []RelayFor
}
type CmdOutput struct {
Relays []*RelayOutput
}
co := CmdOutput{}
enc := json.NewEncoder(w.GetWriter())
if args.Pretty {
enc.SetIndent("", " ")
}
for k, v := range relays {
ro := RelayOutput{NebulaAddr: v.vpnAddrs[0]}
co.Relays = append(co.Relays, &ro)
relayHI := ifce.hostMap.QueryVpnAddr(v.vpnAddrs[0])
if relayHI == nil {
ro.RelayForAddrs = append(ro.RelayForAddrs, RelayFor{Error: "could not find hostinfo"})
continue
}
for _, vpnAddr := range relayHI.relayState.CopyRelayForIps() {
rf := RelayFor{}
r, ok := relayHI.relayState.GetRelayForByAddr(vpnAddr)
if ok {
t := ""
switch r.Type {
case ForwardingType:
t = "forwarding"
case TerminalType:
t = "terminal"
default:
t = "unknown"
}
s := ""
switch r.State {
case Requested:
s = "requested"
case Established:
s = "established"
default:
s = "unknown"
}
rf.LocalIndex = r.LocalIndex
rf.RemoteIndex = r.RemoteIndex
rf.PeerAddr = r.PeerAddr
rf.Type = t
rf.State = s
if rf.LocalIndex != k {
rf.Error = fmt.Sprintf("hostmap LocalIndex '%v' does not match RelayState LocalIndex", k)
}
}
relayedHI := ifce.hostMap.QueryVpnAddr(vpnAddr)
if relayedHI != nil {
rf.RelayedThrough = append(rf.RelayedThrough, relayedHI.relayState.CopyRelayIps()...)
}
ro.RelayForAddrs = append(ro.RelayForAddrs, rf)
}
}
return enc.Encode(co)
}
func printTunnelCommand(ifce *Interface, fs any, a []string, w commands.StringWriter) error {
args, ok := fs.(*printTunnelFlags)
if !ok {
return fmt.Errorf("internal error: expected *printTunnelFlags but was %T", fs)
}
vpnAddr, err := parseVpnAddr(a, "vpn addr", w)
if err != nil {
return err
}
hostInfo := ifce.hostMap.QueryVpnAddr(vpnAddr)
if hostInfo == nil {
return commands.Fail(w, fmt.Sprintf("Could not find tunnel for vpn addr: %v", a[0]))
}
enc := json.NewEncoder(w.GetWriter())
if args.Pretty {
enc.SetIndent("", " ")
}
return enc.Encode(copyHostInfo(hostInfo, ifce.hostMap.GetPreferredRanges()))
}
func deviceInfoCommand(ifce *Interface, fs any, w commands.StringWriter) error {
flags, ok := fs.(*deviceInfoFlags)
if !ok {
return fmt.Errorf("internal error: expected flags to be deviceInfoFlags but was %+v", fs)
}
data := struct {
Name string `json:"name"`
Cidr []netip.Prefix `json:"cidr"`
}{
Name: ifce.inside.Name(),
Cidr: make([]netip.Prefix, len(ifce.inside.Networks())),
}
copy(data.Cidr, ifce.inside.Networks())
if flags.Json || flags.Pretty {
js := json.NewEncoder(w.GetWriter())
if flags.Pretty {
js.SetIndent("", " ")
}
return js.Encode(data)
} else {
return w.WriteLine(fmt.Sprintf("name=%v cidr=%v", data.Name, data.Cidr))
}
}
// reloadCommand reloads the config and writes the result to the client. It reloads even if the client has gone away
func reloadCommand(l *slog.Logger, c *config.C, w commands.StringWriter) error {
writeErr := w.WriteLine("Reloading config")
res := c.ReloadConfig()
res.Log(l)
if err := w.WriteLine(res.String()); writeErr == nil {
writeErr = err
}
if writeErr != nil {
return writeErr
}
if res.Err != nil || res.Failed {
return &commands.Failure{Msg: res.String()}
}
return nil
}
+83
View File
@@ -0,0 +1,83 @@
package commands
import (
"flag"
)
// CommandFlags is called before help or command execution to parse command line flags.
// It should return a flag.FlagSet instance and a pointer to the struct that will contain parsed flags
type CommandFlags func() (*flag.FlagSet, any)
// CommandCallback is called when your command should execute.
// fs will be a pointer to the struct provided by Command.Flags callback, if there was one. -h and -help are reserved
// and handled automatically for you.
// a will be any unconsumed arguments, if no Command.Flags was available this will be all the flags passed in.
// w is the writer to use when sending messages back to the client.
// An error returned by the callback is handed back to whoever dispatched the command, which reports a failure: exit
// status 1 over ssh. The error is not shown to the user. For a failure the user caused or can act on, write the
// message and return it as a *Failure, which Fail does in one step; the surface then reports the failure without
// logging what the user was already told. Any other error is logged by the surface as well.
type CommandCallback func(fs any, a []string, w StringWriter) error
type Command struct {
Name Name
ShortDescription string
Help string
Flags CommandFlags
Callback CommandCallback
}
// Failure is the error a command returns for a failure it has already explained to the user:
// the message it wrote is the whole story. A surface reports it as a failure and does not
// log it.
type Failure struct {
Msg string
}
func (f *Failure) Error() string {
return f.Msg
}
// Fail writes msg to w and returns it as a *Failure, so a command reports a failure to the
// user and to its surface in one line. If the write itself fails, that error is returned
// instead: the user was not told anything.
func Fail(w StringWriter, msg string) error {
if err := w.WriteLine(msg); err != nil {
return err
}
return &Failure{Msg: msg}
}
func execCommand(c *Command, args []string, w StringWriter) error {
var (
fl *flag.FlagSet
fs any
)
if c.Flags != nil {
fl, fs = c.Flags()
if fl != nil {
// SetOutput() here in case fl.Parse dumps usage.
fl.SetOutput(w.GetWriter())
err := fl.Parse(args)
if err != nil {
// fl.Parse has dumped error information to the user via the w writer.
return err
}
args = fl.Args()
}
}
return c.Callback(fs, args, w)
}
func checkHelpArgs(args []string) bool {
for _, a := range args {
if a == "-h" || a == "-help" {
return true
}
}
return false
}
+28
View File
@@ -0,0 +1,28 @@
package commands
// Name identifies a command. Every command nebula serves is a constant here, so a
// surface, a client and a test spell it the same way and are told at compile time when
// they do not. It is a string rather than an integer so a test can register a throwaway
// command without adding it here, and so the wire carries the name as the string it has
// always been.
type Name string
const (
ChangeRemote Name = "change-remote"
CloseTunnel Name = "close-tunnel"
CreateTunnel Name = "create-tunnel"
DeviceInfo Name = "device-info"
Help Name = "help"
ListHostmap Name = "list-hostmap"
ListLighthouseAddrmap Name = "list-lighthouse-addrmap"
ListPendingHostmap Name = "list-pending-hostmap"
LogFormat Name = "log-format"
LogLevel Name = "log-level"
Logout Name = "logout"
PrintCert Name = "print-cert"
PrintRelays Name = "print-relays"
PrintTunnel Name = "print-tunnel"
QueryLighthouse Name = "query-lighthouse"
Reload Name = "reload"
Version Name = "version"
)
+181
View File
@@ -0,0 +1,181 @@
package commands
import (
"cmp"
"fmt"
"maps"
"slices"
"strings"
"sync"
"github.com/anmitsu/go-shlex"
)
// Registry holds a set of commands. New is the only way to make one.
//
// It is safe for concurrent use. Commands may be registered while a surface is serving
// from the registry: the ssh server clones it per session, and a clone may overlap a
// Register.
type Registry struct {
mu sync.RWMutex
commands map[Name]*Command
}
// New returns an empty Registry with only `help` available.
func New() *Registry {
r := &Registry{commands: map[Name]*Command{}}
r.registerHelp()
return r
}
func (r *Registry) registerHelp() {
r.Register(&Command{
Name: Help,
ShortDescription: "prints available commands or help <command> for specific usage info",
Callback: func(a any, args []string, w StringWriter) error {
return r.help(args, w)
},
})
}
// Register adds a command that can be run by a user. A command registered under a name
// already held replaces it; see Clone for what that means for Help.
func (r *Registry) Register(c *Command) {
r.mu.Lock()
defer r.mu.Unlock()
r.commands[c.Name] = c
}
// Get returns the command registered under name, or nil if there is no such command.
func (r *Registry) Get(name Name) *Command {
r.mu.RLock()
defer r.mu.RUnlock()
return r.commands[name]
}
// Match returns the names of every command starting with prefix, sorted, as strings for
// a prompt to complete and print.
func (r *Registry) Match(prefix string) []string {
cmds := make([]string, 0)
r.mu.RLock()
for n := range r.commands {
if strings.HasPrefix(string(n), prefix) {
cmds = append(cmds, string(n))
}
}
r.mu.RUnlock()
slices.Sort(cmds)
return cmds
}
// All returns every registered command sorted by name, which is the order help lists them.
func (r *Registry) All() []*Command {
r.mu.RLock()
defer r.mu.RUnlock()
cmds := slices.Collect(maps.Values(r.commands))
slices.SortFunc(cmds, func(a, b *Command) int { return cmp.Compare(a.Name, b.Name) })
return cmds
}
// Clone returns a copy holding the same commands, with help bound to the copy: a
// command registered on the clone alone is listed and described there, and one the
// clone withholds is not. That rebinding replaces whatever was registered under Help on
// the source, built-in or not.
func (r *Registry) Clone() *Registry {
r.mu.RLock()
c := &Registry{commands: maps.Clone(r.commands)}
r.mu.RUnlock()
c.registerHelp()
return c
}
// Dispatch parses a command line and runs it, reporting problems to w. The error is for
// the caller to log, not to show; see CommandCallback.
func (r *Registry) Dispatch(line string, w StringWriter) error {
// A line shlex cannot split, an unbalanced quote.
args, err := shlex.Split(line, true)
if err != nil {
return Fail(w, "could not parse the command line: "+err.Error())
}
return r.DispatchArgs(args, w)
}
// DispatchArgs runs an already tokenized command line. The error is as for Dispatch.
func (r *Registry) DispatchArgs(args []string, w StringWriter) error {
if len(args) == 0 {
r.dumpCommands(w)
return nil
}
c := r.Get(Name(args[0]))
if c == nil {
// Only the first word was looked up, so only it is named.
msg := fmt.Sprintf("did not understand: %s", args[0])
if err := w.WriteLine(msg); err != nil {
return err
}
r.dumpCommands(w)
return &Failure{Msg: msg}
}
if checkHelpArgs(args) {
// Through the registered help command, so -h and `help <name>` are one path.
return execCommand(r.Get(Help), []string{string(c.Name)}, w)
}
return execCommand(c, args[1:], w)
}
// dumpCommands writes the list of available commands to w.
func (r *Registry) dumpCommands(w StringWriter) {
err := w.WriteLine("Available commands:")
if err != nil {
return
}
// In the order All gives, which is the display order.
cmds := make([]string, 0)
for _, l := range r.All() {
cmds = append(cmds, fmt.Sprintf("%s - %s", l.Name, l.ShortDescription))
}
_ = w.Write(strings.Join(cmds, "\n") + "\n\n")
}
// help writes the command list, or usage for the single command named in a.
func (r *Registry) help(a []string, w StringWriter) error {
// Just typed help
if len(a) == 0 {
r.dumpCommands(w)
return nil
}
// We are printing a specific commands help text
cmd := r.Get(Name(a[0]))
if cmd == nil {
return Fail(w, "Command not available "+a[0])
}
if err := w.WriteLine(fmt.Sprintf("%s - %s", cmd.Name, cmd.ShortDescription)); err != nil {
return err
}
if cmd.Help != "" {
if err := w.WriteLine(fmt.Sprintf(" %s", cmd.Help)); err != nil {
return err
}
}
if cmd.Flags != nil {
fs, _ := cmd.Flags()
if fs != nil {
fs.SetOutput(w.GetWriter())
fs.PrintDefaults()
}
}
return nil
}
+176
View File
@@ -0,0 +1,176 @@
package commands
import (
"bytes"
"errors"
"flag"
"strings"
"testing"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
func testRegistry() *Registry {
r := New()
r.Register(&Command{
Name: "served",
ShortDescription: "a command this surface serves",
Callback: func(fs any, a []string, w StringWriter) error {
return w.WriteLine("read " + strings.Join(a, ","))
},
})
r.Register(&Command{
Name: "withheld",
ShortDescription: "a command this surface withholds",
Callback: func(fs any, a []string, w StringWriter) error {
return w.WriteLine("wrote")
},
})
return r
}
// TestFlagUsageGoesToTheWriter pins that a parse failure's usage lands on the command's
// writer.
func TestFlagUsageGoesToTheWriter(t *testing.T) {
r := New()
r.Register(&Command{
Name: "emit",
ShortDescription: "emits",
Flags: func() (*flag.FlagSet, any) {
fl := flag.NewFlagSet("", flag.ContinueOnError)
fl.Bool("json", false, "emit json")
return fl, nil
},
Callback: func(fs any, a []string, w StringWriter) error { return nil },
})
merged := &bytes.Buffer{}
require.Error(t, r.DispatchArgs([]string{"emit", "-nope"}, NewStringWriter(merged)))
assert.Contains(t, merged.String(), "flag provided but not defined")
}
// TestDispatchReportsUnparseableLine: a line shlex rejects, an unbalanced quote, is a
// failure the user is told about.
func TestDispatchReportsUnparseableLine(t *testing.T) {
out := &bytes.Buffer{}
err := testRegistry().Dispatch(`served "unterminated`, NewStringWriter(out))
var f *Failure
require.ErrorAs(t, err, &f)
assert.Equal(t, f.Msg+"\n", out.String())
assert.Contains(t, out.String(), "could not parse the command line: ")
}
// TestFailTellsTheUserAndTheSurface: Fail writes the message and returns it as a Failure;
// when the write fails, the user heard nothing, so the write error is what comes back.
func TestFailTellsTheUserAndTheSurface(t *testing.T) {
out := &bytes.Buffer{}
err := Fail(NewStringWriter(out), "no such tunnel")
var f *Failure
require.ErrorAs(t, err, &f)
assert.Equal(t, "no such tunnel", f.Msg)
assert.Equal(t, "no such tunnel\n", out.String())
err = Fail(NewStringWriter(failingWriter{}), "no such tunnel")
require.EqualError(t, err, "gone")
assert.NotErrorAs(t, err, &f)
}
type failingWriter struct{}
func (failingWriter) Write([]byte) (int, error) { return 0, errors.New("gone") }
// TestAllSortsByName: registration order does not matter.
func TestAllSortsByName(t *testing.T) {
r := New()
for _, n := range []Name{"zz", Version, "aa", ChangeRemote} {
r.Register(&Command{Name: n})
}
var got []Name
for _, c := range r.All() {
got = append(got, c.Name)
}
assert.Equal(t, []Name{"aa", ChangeRemote, Help, Version, "zz"}, got)
}
// TestHelpFlagAgreesWithHelpCommand pins that -h answers through the registered help
// command, and that help on a clone describes the clone: a command registered only there
// is known to both -h and `help <name>`, and one withheld from it is known to neither.
func TestHelpFlagAgreesWithHelpCommand(t *testing.T) {
r := testRegistry()
c := r.Clone()
c.Register(&Command{
Name: Logout,
ShortDescription: "ends the session",
Callback: func(fs any, a []string, w StringWriter) error { return nil },
})
flag := &bytes.Buffer{}
c.Dispatch("logout -h", NewStringWriter(flag))
cmd := &bytes.Buffer{}
c.Dispatch("help logout", NewStringWriter(cmd))
assert.Equal(t, cmd.String(), flag.String())
assert.Equal(t, "logout - ends the session\n", flag.String())
// The source never learns of it.
src := &bytes.Buffer{}
r.Dispatch("help logout", NewStringWriter(src))
assert.Equal(t, "Command not available logout\n", src.String())
// And a command both hold is described either way.
flag.Reset()
c.Dispatch("served -h", NewStringWriter(flag))
assert.Equal(t, "served - a command this surface serves\n", flag.String())
}
// TestCloneHelpListsTheClone: the command list help prints is the clone's, so a clone
// that withholds a command does not name it, and the source still does.
func TestCloneHelpListsTheClone(t *testing.T) {
r := testRegistry()
c := New()
c.Register(r.Get("served"))
c = c.Clone()
out := &bytes.Buffer{}
c.Dispatch("help", NewStringWriter(out))
assert.Equal(t, "Available commands:\n"+
"help - prints available commands or help <command> for specific usage info\n"+
"served - a command this surface serves\n\n", out.String())
out.Reset()
c.Dispatch("help withheld", NewStringWriter(out))
assert.Equal(t, "Command not available withheld\n", out.String())
out.Reset()
r.Dispatch("help withheld", NewStringWriter(out))
assert.Equal(t, "withheld - a command this surface withholds\n", out.String())
}
// TestDispatchReturnsTheCommandsError: what the callback returns reaches the caller, so
// a surface can log it.
func TestDispatchReturnsTheCommandsError(t *testing.T) {
r := New()
r.Register(&Command{
Name: "fails",
Callback: func(fs any, a []string, w StringWriter) error {
return errors.New("boom")
},
})
out := &bytes.Buffer{}
require.EqualError(t, r.Dispatch("fails", NewStringWriter(out)), "boom")
assert.Empty(t, out.String(), "a command's own error is not shown; the command speaks for itself")
}
// TestUnknownCommandNamesTheWord: only the first word is looked up, so only it is
// reported, followed by the command list, and the line is a failure to the surface.
func TestUnknownCommandNamesTheWord(t *testing.T) {
out := &bytes.Buffer{}
err := testRegistry().Dispatch(`nope "with args"`, NewStringWriter(out))
var f *Failure
require.ErrorAs(t, err, &f)
assert.Equal(t, "did not understand: nope", f.Msg)
assert.True(t, strings.HasPrefix(out.String(), "did not understand: nope\nAvailable commands:\n"), out.String())
}
+8 -1
View File
@@ -1,7 +1,9 @@
package sshd
package commands
import "io"
// StringWriter is how a command talks back to whoever invoked it: one stream, which the
// ssh surface points at its channel.
type StringWriter interface {
WriteLine(string) error
Write(string) error
@@ -13,6 +15,11 @@ type stringWriter struct {
w io.Writer
}
// NewStringWriter returns a StringWriter that writes to w.
func NewStringWriter(w io.Writer) StringWriter {
return &stringWriter{w: w}
}
func (w *stringWriter) WriteLine(s string) error {
return w.Write(s + "\n")
}
+67
View File
@@ -0,0 +1,67 @@
package nebula
import (
"bytes"
"io"
"net/netip"
"testing"
"github.com/slackhq/nebula/commands"
"github.com/slackhq/nebula/logging"
"github.com/slackhq/nebula/test"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
// TestLogFormatRefusalIsShown: a format the logger does not know is reported to the user
// as a failure, rather than returned bare, which over ssh was silence.
func TestLogFormatRefusalIsShown(t *testing.T) {
l := logging.NewLogger(io.Discard)
out := &bytes.Buffer{}
w := commands.NewStringWriter(out)
var f *commands.Failure
require.ErrorAs(t, logFormatCommand(l, nil, []string{"bogus"}, w), &f)
assert.Equal(t, "Unable to set log format: unknown log format `bogus`. possible formats: [text json]\n", out.String())
out.Reset()
require.NoError(t, logFormatCommand(l, nil, []string{"JSON"}, w))
assert.Equal(t, "Log format is: json\n", out.String())
}
// TestPrintRelaysReportsErrorsAsText: the Error field was an error value, which encodes
// as {} and lost the message. It is the message now, and absent when there is none.
func TestPrintRelaysReportsErrorsAsText(t *testing.T) {
l := test.NewLogger()
hm := newHostMap(l)
relay := netip.MustParseAddr("10.0.0.1")
peer := netip.MustParseAddr("10.0.0.2")
// A relay host that is in Relays but not in Hosts, so its hostinfo cannot be found.
hm.Relays[1] = &HostInfo{vpnAddrs: []netip.Addr{relay}}
ifce := &Interface{hostMap: hm}
out := &bytes.Buffer{}
require.NoError(t, printRelaysCommand(ifce, &printTunnelFlags{}, nil, commands.NewStringWriter(out)))
assert.JSONEq(t, `{"Relays":[{"NebulaAddr":"10.0.0.1","RelayForAddrs":[{"Error":"could not find hostinfo","Type":"","State":"","PeerAddr":"","LocalIndex":0,"RemoteIndex":0,"RelayedThrough":null}]}]}`, out.String())
// The same relay host, found, relaying for a peer whose index matches: no Error key.
hi := hm.Relays[1]
r := &Relay{Type: ForwardingType, State: Established, PeerAddr: peer, LocalIndex: 1, RemoteIndex: 2}
hi.relayState = RelayState{
relayForByAddr: map[netip.Addr]*Relay{peer: r},
relayForByIdx: map[uint32]*Relay{r.LocalIndex: r},
}
hm.Hosts[relay] = hi
out.Reset()
require.NoError(t, printRelaysCommand(ifce, &printTunnelFlags{}, nil, commands.NewStringWriter(out)))
assert.JSONEq(t, `{"Relays":[{"NebulaAddr":"10.0.0.1","RelayForAddrs":[{"Type":"forwarding","State":"established","PeerAddr":"10.0.0.2","LocalIndex":1,"RemoteIndex":2,"RelayedThrough":null}]}]}`, out.String())
// And an index mismatch names it.
hm.Relays[7] = hi
delete(hm.Relays, 1)
out.Reset()
require.NoError(t, printRelaysCommand(ifce, &printTunnelFlags{}, nil, commands.NewStringWriter(out)))
assert.Contains(t, out.String(), `"Error":"hostmap LocalIndex '7' does not match RelayState LocalIndex"`)
}
+31 -22
View File
@@ -26,9 +26,12 @@ type C struct {
files []string
Settings map[string]any
oldSettings map[string]any
callbacks []func(*C)
callbacks []func(*C, *Reload)
l *slog.Logger
reloadLock sync.Mutex
// Set by the first reload that needs a restart and never cleared. Guarded by reloadLock
needsRestart bool
}
func NewC(l *slog.Logger) *C {
@@ -72,8 +75,9 @@ func (c *C) LoadString(raw string) error {
// RegisterReloadCallback stores a function to be called when a config reload is triggered. The functions registered
// here should decide if they need to make a change to the current process before making the change. HasChanged can be
// used to help decide if a change is necessary.
// These functions should return quickly or spawn their own go routine if they will take a while
func (c *C) RegisterReloadCallback(f func(*C)) {
// These functions should return quickly or spawn their own go routine if they will take a while.
// Use r to report anything that failed or needs a restart to apply.
func (c *C) RegisterReloadCallback(f func(c *C, r *Reload)) {
c.callbacks = append(c.callbacks, f)
}
@@ -144,50 +148,55 @@ func (c *C) CatchHUP(ctx context.Context) {
return
case <-ch:
c.l.Info("Caught HUP, reloading config")
c.ReloadConfig()
c.ReloadConfig().Log(c.l)
}
}
}()
}
func (c *C) ReloadConfig() {
// ReloadConfig reloads the config from the path provided to Load and runs every registered callback. A config that
// fails to load is logged here, like any other failure, and the result carries the error for the summary.
func (c *C) ReloadConfig() ReloadResult {
c.reloadLock.Lock()
defer c.reloadLock.Unlock()
c.oldSettings = make(map[string]any)
maps.Copy(c.oldSettings, c.Settings)
err := c.Load(c.path)
if err != nil {
c.l.Error("Error occurred while reloading config",
"config_path", c.path,
"error", err,
)
return
if err := c.Load(c.path); err != nil {
c.l.Error("Error occurred while reloading config", "config_path", c.path, "error", err)
return ReloadResult{Err: err, RestartRequired: c.needsRestart}
}
for _, v := range c.callbacks {
v(c)
}
return c.runCallbacks()
}
func (c *C) ReloadConfigString(raw string) error {
// ReloadConfigString is ReloadConfig for a config held in memory rather than on disk.
func (c *C) ReloadConfigString(raw string) ReloadResult {
c.reloadLock.Lock()
defer c.reloadLock.Unlock()
c.oldSettings = make(map[string]any)
maps.Copy(c.oldSettings, c.Settings)
err := c.LoadString(raw)
if err != nil {
return err
if err := c.LoadString(raw); err != nil {
c.l.Error("Error occurred while reloading config", "error", err)
return ReloadResult{Err: err, RestartRequired: c.needsRestart}
}
return c.runCallbacks()
}
func (c *C) runCallbacks() ReloadResult {
r := &Reload{c: c}
for _, v := range c.callbacks {
v(c)
v(c, r)
}
return nil
if r.restart {
c.needsRestart = true
}
return ReloadResult{Failed: r.failed, RestartRequired: c.needsRestart, RestartKeys: r.keys}
}
// GetString will get the string for k or return the default d if not found or invalid
@@ -391,7 +400,7 @@ func (c *C) parse() error {
var nm map[string]any
err = yaml.Unmarshal(b, &nm)
if err != nil {
return err
return fmt.Errorf("%s: %w", path, err)
}
// We need to use WithAppendSlice so that firewall rules in separate
+2 -2
View File
@@ -19,7 +19,7 @@ func TestConfig_Load(t *testing.T) {
// invalid yaml
c := NewC(l)
os.WriteFile(filepath.Join(dir, "01.yaml"), []byte(" invalid yaml"), 0644)
require.EqualError(t, c.Load(dir), "yaml: unmarshal errors:\n line 1: cannot unmarshal !!str `invalid...` into map[string]interface {}")
require.EqualError(t, c.Load(dir), filepath.Join(dir, "01.yaml")+": yaml: unmarshal errors:\n line 1: cannot unmarshal !!str `invalid...` into map[string]interface {}")
// simple multi config merge
c = NewC(l)
@@ -129,7 +129,7 @@ func TestConfig_ReloadConfig(t *testing.T) {
os.WriteFile(filepath.Join(dir, "01.yaml"), []byte("outer:\n inner: ho"), 0644)
c.RegisterReloadCallback(func(c *C) {
c.RegisterReloadCallback(func(c *C, _ *Reload) {
done <- true
})
+98
View File
@@ -0,0 +1,98 @@
package config
import (
"context"
"fmt"
"log/slog"
"slices"
)
// ReloadResult describes how a reload went. Problems are logged where they happen, Log writes a one line summary.
type ReloadResult struct {
// Err is set when the new config could not be read or parsed, no callbacks ran
Err error
// Failed is true when a callback could not apply part of the new config
Failed bool
// RestartRequired is true once any reload has needed a restart, and stays true until the process restarts
RestartRequired bool
// RestartKeys are the keys this reload changed that only apply on a restart
RestartKeys []string
}
// Summary returns a one line summary of the reload and the level to log it at.
func (r ReloadResult) Summary() (slog.Level, string) {
switch {
case r.Err != nil && r.RestartRequired:
return slog.LevelError, "Config reload failed, the new config did not load and a restart is still required"
case r.Err != nil:
return slog.LevelError, "Config reload failed, the new config did not load"
case r.Failed && r.RestartRequired:
return slog.LevelError, "Config reload completed with errors and requires a restart to fully apply, review the logs"
case r.Failed:
return slog.LevelError, "Config reload completed with errors, review the logs"
case r.RestartRequired:
return slog.LevelWarn, "Config reload was completed but requires a restart to fully apply"
default:
return slog.LevelInfo, "Config reload complete"
}
}
// String is the summary with Err and RestartKeys included, for output other than the log.
func (r ReloadResult) String() string {
_, msg := r.Summary()
if r.Err != nil {
msg = fmt.Sprintf("%s: %v", msg, r.Err)
}
if len(r.RestartKeys) > 0 {
msg = fmt.Sprintf("%s, keys %v", msg, r.RestartKeys)
}
return msg
}
// Log writes Summary to l.
func (r ReloadResult) Log(l *slog.Logger) {
level, msg := r.Summary()
if len(r.RestartKeys) > 0 {
l.Log(context.Background(), level, msg, "keys", r.RestartKeys)
return
}
l.Log(context.Background(), level, msg)
}
// Reload is passed to each reload callback to report how applying the new config went.
type Reload struct {
c *C
failed bool
restart bool
// Changed keys that only apply on a restart, from every callback
keys []string
}
// Failed marks the reload as failed, the caller logs why. Safe to call on a nil Reload, which is what the initial load
// passes.
func (r *Reload) Failed() {
if r != nil {
r.failed = true
}
}
// RestartRequired marks the reload as needing a restart, the caller logs why. Safe to call on a nil Reload.
func (r *Reload) RestartRequired() {
if r != nil {
r.restart = true
}
}
// RestartRequiredIfChanged marks the reload as needing a restart if any of keys changed, and adds them to RestartKeys.
// Use it for keys that are only read at startup. Safe to call on a nil Reload.
func (r *Reload) RestartRequiredIfChanged(keys ...string) {
if r == nil {
return
}
for _, k := range keys {
if r.c.HasChanged(k) && !slices.Contains(r.keys, k) {
r.keys = append(r.keys, k)
r.restart = true
}
}
}
+108
View File
@@ -0,0 +1,108 @@
package config
import (
"errors"
"log/slog"
"testing"
"github.com/slackhq/nebula/test"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
func TestReloadConfigString(t *testing.T) {
c := NewC(test.NewLogger())
require.NoError(t, c.LoadString("a: 1"))
fail := false
ran := 0
c.RegisterReloadCallback(func(c *C, r *Reload) {
ran++
if fail {
r.Failed()
}
})
assert.Equal(t, ReloadResult{}, c.ReloadConfigString("a: 2"))
fail = true
assert.Equal(t, ReloadResult{Failed: true}, c.ReloadConfigString("a: 3"))
assert.Equal(t, 2, ran)
// A config that fails to load runs no callbacks and returns the error
res := c.ReloadConfigString("a: [")
require.Error(t, res.Err)
assert.False(t, res.Failed)
assert.Equal(t, 2, ran)
assert.Equal(t, 3, c.GetInt("a", 0))
}
func TestReloadRestartRequiredLatches(t *testing.T) {
c := NewC(test.NewLogger())
require.NoError(t, c.LoadString("a: 1\nb: 1"))
c.RegisterReloadCallback(func(c *C, r *Reload) { r.RestartRequiredIfChanged("a", "b") })
// Nothing that needs a restart changed
assert.Equal(t, ReloadResult{}, c.ReloadConfigString("a: 1\nb: 1\nc: 1"))
assert.Equal(t, ReloadResult{RestartRequired: true, RestartKeys: []string{"a"}}, c.ReloadConfigString("a: 2\nb: 1\nc: 1"))
// The restart sticks through a reload with no changes, reverting the key and a config that fails to load. Only the
// reload that changed the key lists it
assert.Equal(t, ReloadResult{RestartRequired: true}, c.ReloadConfigString("a: 2\nb: 1\nc: 1"))
assert.Equal(t, ReloadResult{RestartRequired: true, RestartKeys: []string{"a"}}, c.ReloadConfigString("a: 1\nb: 1\nc: 1"))
res := c.ReloadConfigString("a: [")
require.Error(t, res.Err)
assert.True(t, res.RestartRequired)
}
// A key reported by more than one callback is listed once
func TestReloadRestartKeysDeduped(t *testing.T) {
c := NewC(test.NewLogger())
require.NoError(t, c.LoadString("a: 1\nb: 1"))
c.RegisterReloadCallback(func(c *C, r *Reload) { r.RestartRequiredIfChanged("a") })
c.RegisterReloadCallback(func(c *C, r *Reload) { r.RestartRequiredIfChanged("a", "b") })
assert.Equal(t, []string{"a", "b"}, c.ReloadConfigString("a: 2\nb: 2").RestartKeys)
}
// A callback can require a restart without a key changing, a file the config points at may have
func TestReloadRestartRequiredWithoutAChange(t *testing.T) {
c := NewC(test.NewLogger())
require.NoError(t, c.LoadString("a: 1"))
c.RegisterReloadCallback(func(c *C, r *Reload) { r.RestartRequired() })
assert.Equal(t, ReloadResult{RestartRequired: true}, c.ReloadConfigString("a: 1"))
}
func TestReloadResultSummary(t *testing.T) {
cases := []struct {
res ReloadResult
level slog.Level
msg string
}{
{ReloadResult{}, slog.LevelInfo, "Config reload complete"},
{ReloadResult{RestartRequired: true}, slog.LevelWarn, "Config reload was completed but requires a restart to fully apply"},
{ReloadResult{Failed: true}, slog.LevelError, "Config reload completed with errors, review the logs"},
{ReloadResult{Failed: true, RestartRequired: true}, slog.LevelError, "Config reload completed with errors and requires a restart to fully apply, review the logs"},
{ReloadResult{Err: errors.New("bad yaml")}, slog.LevelError, "Config reload failed, the new config did not load"},
{ReloadResult{Err: errors.New("bad yaml"), RestartRequired: true}, slog.LevelError, "Config reload failed, the new config did not load and a restart is still required"},
}
for _, tc := range cases {
level, msg := tc.res.Summary()
assert.Equal(t, tc.level, level, tc.msg)
assert.Equal(t, tc.msg, msg)
}
res := ReloadResult{RestartRequired: true, RestartKeys: []string{"a", "b"}}
assert.Equal(t, "Config reload was completed but requires a restart to fully apply, keys [a b]", res.String())
// String includes the error
assert.Equal(t, "Config reload failed, the new config did not load: bad yaml", ReloadResult{Err: errors.New("bad yaml")}.String())
}
// The initial load passes a nil Reload
func TestReloadNilIsSafe(t *testing.T) {
var r *Reload
r.Failed()
r.RestartRequired()
r.RestartRequiredIfChanged("a")
}
+4 -5
View File
@@ -57,7 +57,8 @@ func newConnectionManagerFromConfig(l *slog.Logger, c *config.C, hm *HostMap, p
}
cm.reload(c, true)
c.RegisterReloadCallback(func(c *config.C) {
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
r.RestartRequiredIfChanged("timers.connection_alive_interval", "timers.pending_deletion_interval")
cm.reload(c, false)
})
@@ -174,10 +175,8 @@ func (cm *connectionManager) doTrafficCheck(localIndex uint32, p, nb, out []byte
switch decision {
case deleteTunnel:
if cm.hostMap.DeleteHostInfo(hostinfo) {
// Only clearing the lighthouse cache if this is the last hostinfo for this vpn ip in the hostmap
cm.intf.lightHouse.DeleteVpnAddrs(hostinfo.vpnAddrs)
}
// Local delete only, clearing the lighthouse cache if this was the last hostinfo for the vpn addr
cm.intf.closeTunnel(hostinfo)
case closeTunnel:
cm.intf.sendCloseTunnel(hostinfo)
+3
View File
@@ -7,6 +7,7 @@ import (
"testing"
"time"
"github.com/rcrowley/go-metrics"
"github.com/slackhq/nebula/cert"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/overlay/overlaytest"
@@ -21,6 +22,8 @@ func newTestLighthouse() *LightHouse {
l: test.NewLogger(),
addrMap: map[netip.Addr]*RemoteList{},
queryChan: make(chan netip.Addr, 10),
// Nothing drains queryChan here, so an 11th query is dropped.
queryDropped: metrics.NilCounter{},
}
lighthouses := []netip.Addr{}
staticList := map[netip.Addr]struct{}{}
+119
View File
@@ -2,6 +2,8 @@ package nebula
import (
"net/netip"
"strconv"
"sync"
"testing"
"time"
@@ -11,6 +13,7 @@ import (
ct "github.com/slackhq/nebula/cert_test"
"github.com/slackhq/nebula/handshake"
"github.com/slackhq/nebula/header"
"github.com/slackhq/nebula/overlay"
"github.com/slackhq/nebula/test"
"github.com/slackhq/nebula/udp"
"github.com/stretchr/testify/assert"
@@ -147,6 +150,122 @@ func TestSendNoMetricsCloseTunnelKeepsRebindEpoch(t *testing.T) {
assert.True(t, hostinfo.markOut(1), "a CloseTunnel send must not consume the rebind epoch")
}
// capturingConn is a udp.Conn that records every WriteTo.
type capturingConn struct {
udp.NoopConn
writes [][]byte
addrs []netip.AddrPort
}
func (c *capturingConn) WriteTo(b []byte, addr netip.AddrPort) error {
c.writes = append(c.writes, append([]byte(nil), b...))
c.addrs = append(c.addrs, addr)
return nil
}
// TestSendNoMetricsViaRelay sends through a relay from the mtu sized buffer the cached packet flush uses. Payloads up
// to overlay.MaxMTU go out intact, anything bigger is dropped rather than outgrowing the buffer, which once panicked.
func TestSendNoMetricsViaRelay(t *testing.T) {
tests := []struct {
n int
sent bool
}{
{0, true},
{1300, true},
{overlay.MaxMTU, true},
{overlay.MaxMTU + 1, false}, // too big for the relay's tag
{mtu - 47, false}, // too big for the inner tag, this panicked
{9000, false}, // a full packet at tun.mtu 9000
}
for _, tt := range tests {
t.Run(strconv.Itoa(tt.n), func(t *testing.T) {
// One tunnel to the target, carried end to end, and one to the relay that wraps it
targetInit, targetResp := runTestHandshake(t)
relayInit, relayResp := runTestHandshake(t)
ci, err := newConnectionStateFromResult(targetInit)
require.NoError(t, err)
relayCI, err := newConnectionStateFromResult(relayInit)
require.NoError(t, err)
targetAddr := netip.MustParseAddr("10.0.0.2")
relayAddr := netip.MustParseAddr("10.0.0.3")
relayRemote := netip.MustParseAddrPort("192.0.2.1:4242")
relay := &Relay{Type: TerminalType, State: Established, LocalIndex: 5, RemoteIndex: 6, PeerAddr: targetAddr}
relayHI := &HostInfo{
vpnAddrs: []netip.Addr{relayAddr},
ConnectionState: relayCI,
relayState: RelayState{
relayForByAddr: map[netip.Addr]*Relay{targetAddr: relay},
relayForByIdx: map[uint32]*Relay{relay.LocalIndex: relay},
},
}
relayHI.remote.Store(&relayRemote)
// No remote of its own, so the send has to go via the relay
hostinfo := &HostInfo{
vpnAddrs: []netip.Addr{targetAddr},
ConnectionState: ci,
remoteIndexId: 7,
relayState: RelayState{
relays: []netip.Addr{relayAddr},
relayForByAddr: map[netip.Addr]*Relay{},
relayForByIdx: map[uint32]*Relay{},
},
}
l := test.NewLogger()
hm := newHostMap(l)
hm.Hosts[relayAddr] = relayHI
conn := &capturingConn{}
f := &Interface{
l: l,
hostMap: hm,
messageMetrics: &MessageMetrics{txExhausted: metrics.NewCounter()},
writers: []udp.Conn{conn},
}
f.connectionManager = &connectionManager{intf: f, relayUsed: map[uint32]struct{}{}, relayUsedLock: &sync.RWMutex{}}
payload := make([]byte, tt.n)
for i := range payload {
payload[i] = byte(i)
}
counter := ci.messageCounter.Load()
f.sendNoMetrics(header.Message, 0, ci, hostinfo, netip.AddrPort{}, payload, make([]byte, 12), make([]byte, mtu), 0)
if !tt.sent {
assert.Empty(t, conn.writes)
assert.Equal(t, counter, ci.messageCounter.Load(), "a dropped packet must not spend a counter")
return
}
require.Len(t, conn.writes, 1)
assert.Equal(t, relayRemote, conn.addrs[0])
pkt := conn.writes[0]
// The relay authenticates the whole packet under its own header
outer := &header.H{}
require.NoError(t, outer.Parse(pkt))
assert.Equal(t, header.MessageRelay, outer.Subtype)
assert.Equal(t, relay.RemoteIndex, outer.RemoteIndex)
relayPeer, err := newConnectionStateFromResult(relayResp)
require.NoError(t, err)
require.NoError(t, relayPeer.VerifyRelay(l, outer.MessageCounter, pkt, make([]byte, 12)))
// Inside it, the target decrypts the end to end packet back to the payload
inner := pkt[header.Len : len(pkt)-relayPeer.dKey.Overhead()]
h := &header.H{}
require.NoError(t, h.Parse(inner))
assert.Equal(t, header.Message, h.Type)
assert.Equal(t, hostinfo.remoteIndexId, h.RemoteIndex)
targetPeer, err := newConnectionStateFromResult(targetResp)
require.NoError(t, err)
got, err := targetPeer.Decrypt(l, h.MessageCounter, inner, make([]byte, 12))
require.NoError(t, err)
assert.Equal(t, payload, got)
})
}
}
func TestNewConnectionStateFromResult(t *testing.T) {
initR, respR := runTestHandshake(t)
+9
View File
@@ -11,6 +11,7 @@ import (
"syscall"
"github.com/slackhq/nebula/cert"
"github.com/slackhq/nebula/commands"
"github.com/slackhq/nebula/header"
"github.com/slackhq/nebula/overlay"
)
@@ -47,6 +48,7 @@ type Control struct {
f *Interface
l *slog.Logger
commands *commands.Registry
ctx context.Context
cancel context.CancelFunc
sshStart func()
@@ -372,6 +374,13 @@ func (c *Control) CloseAllTunnels(excludeLighthouses bool) (closed int) {
return
}
// Commands returns the registry the daemon's surfaces serve from. It is the live
// registry, not a copy: a command registered on it is served to every ssh session that
// starts afterwards. Clone it to run commands without that reach.
func (c *Control) Commands() *commands.Registry {
return c.commands
}
func (c *Control) Device() overlay.Device {
return c.f.inside
}
+1 -1
View File
@@ -151,7 +151,7 @@ func (c *fakeConn) WriteTo(_ []byte, _ netip.AddrPort) error { return nil }
func (c *fakeConn) WriteBatch(bufs [][]byte, _ []netip.AddrPort) (int, error) {
return len(bufs), nil
}
func (c *fakeConn) ReloadConfig(_ *config.C) {}
func (c *fakeConn) ReloadConfig(_ *config.C, _ *config.Reload) {}
func (c *fakeConn) SupportsMultipleReaders() bool { return true }
func (c *fakeConn) Close() error { c.closed = true; return nil }
+51 -23
View File
@@ -38,8 +38,8 @@ type dnsServer struct {
serverMu sync.Mutex
server *dns.Server
// started is closed once `server` has finished binding (or after
// ListenAndServe returns on a bind failure). Stop waits on it before
// started is closed once `server` is serving (or after serve returns
// without ever starting). Stop waits on it before
// calling Shutdown to avoid the miekg/dns "server not started" race
// where a Shutdown that arrives before bind completes is silently
// ignored, leaving the listener running forever.
@@ -69,9 +69,10 @@ func newDnsServerFromConfig(ctx context.Context, l *slog.Logger, pki *PKI, hostM
ds.mux = dns.NewServeMux()
ds.mux.HandleFunc(".", ds.handleDnsRequest)
c.RegisterReloadCallback(func(c *config.C) {
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
if err := ds.reload(c, false); err != nil {
ds.l.Error("Failed to reload DNS responder from config", "error", err)
r.Failed()
}
})
@@ -83,7 +84,7 @@ func newDnsServerFromConfig(ctx context.Context, l *slog.Logger, pki *PKI, hostM
}
// reload applies the latest config and reconciles the running state with it:
// - enabled toggled on -> spawn a runner
// - enabled toggled on -> bind, then serve in the background
// - enabled toggled off -> stop the runner
// - listen address changed (while running) -> restart on the new address
// - everything else -> no-op
@@ -120,13 +121,20 @@ func (d *dnsServer) reload(c *config.C, initial bool) error {
return nil
}
if !running {
// Was disabled (or never started); bring it up now.
go d.Start()
} else if !sameAddr {
// Stop clears the slot before shutting down, otherwise the Start below can find the dying server and refuse
if running && !sameAddr {
// Stop clears the slot before shutting down, otherwise the listen below can find the dying server and refuse
d.Stop()
go d.Start()
running = false
}
if !running {
// Bind here rather than in the background, so a reload that can't listen says so
server, started, err := d.listen()
if err != nil {
return err
}
if server != nil {
go d.serve(server, started)
}
}
// Refresh the self entry every enabled reload so cert renewals that change our name or VPN addresses are picked up.
@@ -155,30 +163,50 @@ func (d *dnsServer) shutdownServer(srv *dns.Server, started chan struct{}, reaso
// Must be invoked after the tun device is active so that lighthouse.dns.host
// may bind to a nebula IP.
func (d *dnsServer) Start() {
if !d.enabled.Load() {
server, started, err := d.listen()
if err != nil {
d.l.Warn("Failed to run the DNS responder", "error", err)
return
}
if server != nil {
d.serve(server, started)
}
}
// listen binds the responder and claims the server slot for serve. It returns a nil server, and no error, when DNS is
// disabled or a server is already running.
func (d *dnsServer) listen() (*dns.Server, chan struct{}, error) {
if !d.enabled.Load() {
return nil, nil, nil
}
started := make(chan struct{})
d.serverMu.Lock()
defer d.serverMu.Unlock()
// Re-check enabled under the lock, a disable that raced our check above snapshots the slot under it too.
// Two reloads in quick succession can both spawn a Start, the loser would orphan the live listener past Stop
// A reload and Control.Start can both get here, the loser would orphan the live listener past Stop
if d.ctx.Err() != nil || d.server != nil || !d.enabled.Load() {
d.serverMu.Unlock()
return
return nil, nil, nil
}
pc, err := net.ListenPacket("udp", d.addr)
if err != nil {
return nil, nil, err
}
addr := d.addr
server := &dns.Server{
Addr: addr,
Net: "udp",
Addr: d.addr,
PacketConn: pc,
Handler: d.mux,
NotifyStartedFunc: func() { close(started) },
}
d.server = server
d.started = started
d.serverMu.Unlock()
d.l.Info("Starting DNS responder", "dnsListener", pc.LocalAddr())
return server, started, nil
}
// Per-invocation ctx watcher. Exits when Start does, so we don't leak a
// serve answers queries on a server from listen until Stop or the context ends it.
func (d *dnsServer) serve(server *dns.Server, started chan struct{}) {
// Per-invocation ctx watcher. Exits when serve does, so we don't leak a
// watcher per reload-driven restart.
done := make(chan struct{})
go func() {
@@ -189,16 +217,16 @@ func (d *dnsServer) Start() {
}
}()
d.l.Info("Starting DNS responder", "dnsListener", addr)
err := server.ListenAndServe()
err := server.ActivateAndServe()
close(done)
// If the listener never bound (bind error) NotifyStartedFunc never fires,
// so close started here to release any Stop caller waiting on it.
// If serving failed before it started NotifyStartedFunc never fires, so close started here to release any Stop
// caller waiting on it. ActivateAndServe doesn't close the socket on that path either
select {
case <-started:
default:
close(started)
_ = server.PacketConn.Close()
}
// Release our slot, unless a reload already replaced us, so a dead listener can't block a future Start
+19
View File
@@ -629,3 +629,22 @@ func TestDnsServer_Start_refusesWhenDisabledUnderLock(t *testing.T) {
require.NoError(t, err, "an orphaned listener is still holding the port")
_ = pc.Close()
}
// A reload that can't bind says so, rather than finding out in the background
func TestDnsServer_reload_busyPortFails(t *testing.T) {
busy, err := net.ListenPacket("udp", "127.0.0.1:0")
require.NoError(t, err)
t.Cleanup(func() { _ = busy.Close() })
_, port, err := net.SplitHostPort(busy.LocalAddr().String())
require.NoError(t, err)
ds, c := newTestDnsServer(t)
setDnsConfig(c, "127.0.0.1", port, false, true)
require.NoError(t, ds.reload(c, true))
setDnsConfig(c, "127.0.0.1", port, true, true)
require.Error(t, ds.reload(c, false))
ds.serverMu.Lock()
assert.Nil(t, ds.server, "a failed bind kept the slot")
ds.serverMu.Unlock()
}
+164
View File
@@ -0,0 +1,164 @@
//go:build e2e_testing
package e2e
import (
"bytes"
"fmt"
"path/filepath"
"regexp"
"strings"
"testing"
"time"
"github.com/slackhq/nebula/cert"
"github.com/slackhq/nebula/cert_test"
"github.com/slackhq/nebula/commands"
"github.com/slackhq/nebula/sshd"
"github.com/slackhq/nebula/test"
"github.com/stretchr/testify/require"
)
// volatile lists everything in command output that legitimately differs between runs.
// Each run mints fresh certificates and indexes, so these are replaced with a marker
// before comparison. Keep the list short: anything added here is something the golden
// stops guarding.
var volatile = []struct {
re *regexp.Regexp
with string
}{
{regexp.MustCompile(`[0-9a-f]{40,}`), "<hex>"}, // fingerprints, keys, signatures
{regexp.MustCompile(`\d{4}-\d{2}-\d{2}T\d{2}:\d{2}:\d{2}(?:\.\d+)?(?:Z|[+-]\d{2}:\d{2})`), "<time>"}, // RFC 3339
{regexp.MustCompile(`\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2}(?:\.\d+)? [+-]\d{4} \w+`), "<time>"}, // time.Time.String()
{regexp.MustCompile(`(?s)(-----BEGIN [A-Z ]+-----\n).*?(\n-----END [A-Z ]+-----)`), "${1}<pem>${2}"}, // print-cert -raw
{regexp.MustCompile(`(-----BEGIN [A-Z ]+-----\\n).*?(\\n-----END [A-Z ]+-----)`), "${1}<pem>${2}"}, // the same, inside a JSON string
{regexp.MustCompile(`"(localIndex|remoteIndex|messageCounter)":\d+`), `"$1":<n>`},
{regexp.MustCompile(`(?m)^(\s*"?(?:LocalIndex|RemoteIndex|localIndex|remoteIndex|messageCounter)"?:?\s*)\d+`), "${1}<n>"}, // and the -pretty form
}
func normalise(s string) string {
for _, v := range volatile {
s = v.re.ReplaceAllString(s, v.with)
}
return s
}
// outputCase is one command line to run through the ssh dispatch path.
type outputCase struct {
name string
args []string
}
// TestCommandOutput pins the bytes every command produces, in every format, through the
// ssh dispatch path. It is the safety net for moving rendering out of the commands: a
// change here is a change someone at an ssh prompt would see. Regenerate with:
//
// go test -tags e2e_testing ./e2e -run TestCommandOutput -update
func TestCommandOutput(t *testing.T) {
ca, _, caKey, _ := cert_test.NewTestCaCert(
cert.Version1, cert.Curve_CURVE25519,
time.Now(), time.Now().Add(10*time.Minute),
nil, nil, []string{},
)
control, _, _, _ := newSimpleServer(cert.Version1, ca, caKey, "output-golden", "10.222.0.1/24", nil)
control.Start()
t.Cleanup(func() { control.Stop() })
// As an ssh session does: a clone of the daemon's registry with logout added, which
// help on the clone lists and describes. There is no session to end here.
reg := control.Commands().Clone()
reg.Register(sshd.LogoutCommand(func() {}))
// run dispatches as the ssh exec path does and returns the output and the error, which
// over ssh is exit status 1.
run := func(args ...string) (string, error) {
buf := &bytes.Buffer{}
err := reg.DispatchArgs(args, commands.NewStringWriter(buf))
return buf.String(), err
}
// Seed a pending tunnel so the listing and per-address commands have something to show.
seed, err := run("create-tunnel", "10.222.0.9")
require.NoError(t, err)
require.Equal(t, "Created\n", seed, "seeding a tunnel should succeed before anything else is recorded")
cases := []outputCase{
{"version", nil},
{"device-info", nil},
{"list-hostmap", nil},
{"list-hostmap", []string{"-by-index"}},
{"list-pending-hostmap", nil},
{"list-pending-hostmap", []string{"-by-index"}},
{"list-lighthouse-addrmap", nil},
{"print-cert", nil},
{"print-cert", []string{"-raw"}},
{"print-cert", []string{"10.222.0.9"}},
{"print-cert", []string{"10.222.0.250"}},
{"print-cert", []string{"not-an-address"}},
{"print-tunnel", nil},
{"print-tunnel", []string{"10.222.0.9"}},
{"print-tunnel", []string{"10.222.0.250"}},
{"print-relays", nil},
{"query-lighthouse", nil},
{"query-lighthouse", []string{"10.222.0.9"}},
{"query-lighthouse", []string{"not-an-address"}},
// The test logger cannot change level or format, so these cases record the
// refusal each command prints on it, not a successful change.
{"log-level", nil},
{"log-level", []string{"debug"}},
{"log-level", []string{"bogus"}},
{"log-level", []string{"info"}},
{"log-format", nil},
{"log-format", []string{"json"}},
{"log-format", []string{"text"}},
{"change-remote", nil},
{"change-remote", []string{"10.222.0.9"}},
{"change-remote", []string{"-address", "10.94.0.10:4242", "10.222.0.9"}},
{"change-remote", []string{"-address", "10.94.0.10:4242", "10.222.0.250"}},
{"create-tunnel", nil},
{"create-tunnel", []string{"10.222.0.9"}},
// A fresh address, so this exercises the handshake start with a remote attached
// rather than the "already handshaking" path above.
{"create-tunnel", []string{"-address", "10.94.0.10:4242", "10.222.0.10"}},
{"create-tunnel", []string{"not-an-address"}},
{"close-tunnel", nil},
{"close-tunnel", []string{"10.222.0.250"}},
{"close-tunnel", []string{"not-an-address"}},
{"help", nil},
{"help", []string{"print-cert"}},
{"help", []string{"logout"}},
{"logout", []string{"-h"}},
{"not-a-command", []string{"10.222.0.9"}},
// reload is last so that whatever a reload does to the daemon cannot leak into an
// earlier case. In this harness it fails: the config was loaded from a string and
// has no path to re-read, which is what the recorded result says.
{"reload", nil},
}
got := &strings.Builder{}
for _, c := range cases {
for _, fmtFlags := range [][]string{nil, {"-json"}, {"-pretty"}} {
args := append([]string{c.name}, fmtFlags...)
args = append(args, c.args...)
fmt.Fprintf(got, "$ %s\n", strings.Join(args, " "))
out, err := run(args...)
out = normalise(out)
got.WriteString(out)
// Output that does not end in a newline is recorded as such, the way diff marks
// it, so the golden sees that byte too and the next prompt still starts a line.
if out != "" && !strings.HasSuffix(out, "\n") {
got.WriteString("\\ No newline at end of output\n")
}
// A failed command is recorded as the exit status an ssh caller sees, so the
// golden pins which outcomes count as failures.
if err != nil {
got.WriteString("exit status 1\n")
}
got.WriteString("\n")
}
}
test.Golden(t, filepath.Join("testdata", "commands_output.golden"), got.String())
}
+1 -1
View File
@@ -1513,7 +1513,7 @@ func TestLighthouseUpdateOnReload(t *testing.T) {
// Reload the config. The lighthouse.hosts change triggers TriggerUpdate,
// which wakes the update worker. It calls SendUpdate, initiating a
// handshake to the new lighthouse and caching the HostUpdateNotification.
require.NoError(t, myConfig.ReloadConfigString(string(newCfg)))
require.NoError(t, myConfig.ReloadConfigString(string(newCfg)).Err)
// Route until the lighthouse receives the HostUpdateNotification.
// This covers: handshake stage 1, stage 2, then the cached update.
+720
View File
@@ -0,0 +1,720 @@
$ version
e2e-test
$ version -json
e2e-test
$ version -pretty
e2e-test
$ device-info
name= cidr=[10.222.0.1/24]
$ device-info -json
{"name":"","cidr":["10.222.0.1/24"]}
$ device-info -pretty
{
"name": "",
"cidr": [
"10.222.0.1/24"
]
}
$ list-hostmap
$ list-hostmap -json
[]
$ list-hostmap -pretty
[]
$ list-hostmap -by-index
$ list-hostmap -json -by-index
[]
$ list-hostmap -pretty -by-index
[]
$ list-pending-hostmap
[10.222.0.9]: []
$ list-pending-hostmap -json
[{"vpnAddrs":["10.222.0.9"],"localIndex":<n>,"remoteIndex":<n>,"remoteAddrs":[],"cert":null,"messageCounter":<n>,"currentRemote":"","currentRelaysToMe":[],"currentRelaysThroughMe":[]}]
$ list-pending-hostmap -pretty
[
{
"vpnAddrs": [
"10.222.0.9"
],
"localIndex": <n>,
"remoteIndex": <n>,
"remoteAddrs": [],
"cert": null,
"messageCounter": <n>,
"currentRemote": "",
"currentRelaysToMe": [],
"currentRelaysThroughMe": []
}
]
$ list-pending-hostmap -by-index
$ list-pending-hostmap -json -by-index
[]
$ list-pending-hostmap -pretty -by-index
[]
$ list-lighthouse-addrmap
10.222.0.9: {}
$ list-lighthouse-addrmap -json
[{"vpnAddr":"10.222.0.9","addrs":{}}]
$ list-lighthouse-addrmap -pretty
[
{
"vpnAddr": "10.222.0.9",
"addrs": {}
}
]
$ print-cert
{
"details": {
"curve": "CURVE25519",
"groups": [],
"isCa": false,
"issuer": "<hex>",
"name": "output-golden",
"networks": [
"10.222.0.1/24"
],
"notAfter": "<time>",
"notBefore": "<time>",
"publicKey": "<hex>",
"unsafeNetworks": []
},
"fingerprint": "<hex>",
"signature": "<hex>",
"version": 1
}
$ print-cert -json
{"details":{"curve":"CURVE25519","groups":[],"isCa":false,"issuer":"<hex>","name":"output-golden","networks":["10.222.0.1/24"],"notAfter":"<time>","notBefore":"<time>","publicKey":"<hex>","unsafeNetworks":[]},"fingerprint":"<hex>","signature":"<hex>","version":1}\ No newline at end of output
$ print-cert -pretty
{
"details": {
"curve": "CURVE25519",
"groups": [],
"isCa": false,
"issuer": "<hex>",
"name": "output-golden",
"networks": [
"10.222.0.1/24"
],
"notAfter": "<time>",
"notBefore": "<time>",
"publicKey": "<hex>",
"unsafeNetworks": []
},
"fingerprint": "<hex>",
"signature": "<hex>",
"version": 1
}\ No newline at end of output
$ print-cert -raw
-----BEGIN NEBULA CERTIFICATE-----
<pem>
-----END NEBULA CERTIFICATE-----
$ print-cert -json -raw
{"details":{"curve":"CURVE25519","groups":[],"isCa":false,"issuer":"<hex>","name":"output-golden","networks":["10.222.0.1/24"],"notAfter":"<time>","notBefore":"<time>","publicKey":"<hex>","unsafeNetworks":[]},"fingerprint":"<hex>","signature":"<hex>","version":1}\ No newline at end of output
$ print-cert -pretty -raw
{
"details": {
"curve": "CURVE25519",
"groups": [],
"isCa": false,
"issuer": "<hex>",
"name": "output-golden",
"networks": [
"10.222.0.1/24"
],
"notAfter": "<time>",
"notBefore": "<time>",
"publicKey": "<hex>",
"unsafeNetworks": []
},
"fingerprint": "<hex>",
"signature": "<hex>",
"version": 1
}\ No newline at end of output
$ print-cert 10.222.0.9
Could not find tunnel for vpn addr: 10.222.0.9
exit status 1
$ print-cert -json 10.222.0.9
Could not find tunnel for vpn addr: 10.222.0.9
exit status 1
$ print-cert -pretty 10.222.0.9
Could not find tunnel for vpn addr: 10.222.0.9
exit status 1
$ print-cert 10.222.0.250
Could not find tunnel for vpn addr: 10.222.0.250
exit status 1
$ print-cert -json 10.222.0.250
Could not find tunnel for vpn addr: 10.222.0.250
exit status 1
$ print-cert -pretty 10.222.0.250
Could not find tunnel for vpn addr: 10.222.0.250
exit status 1
$ print-cert not-an-address
The provided vpn addr could not be parsed: not-an-address
exit status 1
$ print-cert -json not-an-address
The provided vpn addr could not be parsed: not-an-address
exit status 1
$ print-cert -pretty not-an-address
The provided vpn addr could not be parsed: not-an-address
exit status 1
$ print-tunnel
No vpn address was provided
exit status 1
$ print-tunnel -json
flag provided but not defined: -json
Usage:
-pretty
pretty prints json
exit status 1
$ print-tunnel -pretty
No vpn address was provided
exit status 1
$ print-tunnel 10.222.0.9
Could not find tunnel for vpn addr: 10.222.0.9
exit status 1
$ print-tunnel -json 10.222.0.9
flag provided but not defined: -json
Usage:
-pretty
pretty prints json
exit status 1
$ print-tunnel -pretty 10.222.0.9
Could not find tunnel for vpn addr: 10.222.0.9
exit status 1
$ print-tunnel 10.222.0.250
Could not find tunnel for vpn addr: 10.222.0.250
exit status 1
$ print-tunnel -json 10.222.0.250
flag provided but not defined: -json
Usage:
-pretty
pretty prints json
exit status 1
$ print-tunnel -pretty 10.222.0.250
Could not find tunnel for vpn addr: 10.222.0.250
exit status 1
$ print-relays
{"Relays":null}
$ print-relays -json
flag provided but not defined: -json
Usage:
-pretty
pretty prints json
exit status 1
$ print-relays -pretty
{
"Relays": null
}
$ query-lighthouse
No vpn address was provided
exit status 1
$ query-lighthouse -json
The provided vpn address could not be parsed: -json
exit status 1
$ query-lighthouse -pretty
The provided vpn address could not be parsed: -pretty
exit status 1
$ query-lighthouse 10.222.0.9
{}
$ query-lighthouse -json 10.222.0.9
The provided vpn address could not be parsed: -json
exit status 1
$ query-lighthouse -pretty 10.222.0.9
The provided vpn address could not be parsed: -pretty
exit status 1
$ query-lighthouse not-an-address
The provided vpn address could not be parsed: not-an-address
exit status 1
$ query-lighthouse -json not-an-address
The provided vpn address could not be parsed: -json
exit status 1
$ query-lighthouse -pretty not-an-address
The provided vpn address could not be parsed: -pretty
exit status 1
$ log-level
Log level is not reconfigurable on this logger
exit status 1
$ log-level -json
Log level is not reconfigurable on this logger
exit status 1
$ log-level -pretty
Log level is not reconfigurable on this logger
exit status 1
$ log-level debug
Log level is not reconfigurable on this logger
exit status 1
$ log-level -json debug
Log level is not reconfigurable on this logger
exit status 1
$ log-level -pretty debug
Log level is not reconfigurable on this logger
exit status 1
$ log-level bogus
Log level is not reconfigurable on this logger
exit status 1
$ log-level -json bogus
Log level is not reconfigurable on this logger
exit status 1
$ log-level -pretty bogus
Log level is not reconfigurable on this logger
exit status 1
$ log-level info
Log level is not reconfigurable on this logger
exit status 1
$ log-level -json info
Log level is not reconfigurable on this logger
exit status 1
$ log-level -pretty info
Log level is not reconfigurable on this logger
exit status 1
$ log-format
Log format is not reconfigurable on this logger
exit status 1
$ log-format -json
Log format is not reconfigurable on this logger
exit status 1
$ log-format -pretty
Log format is not reconfigurable on this logger
exit status 1
$ log-format json
Log format is not reconfigurable on this logger
exit status 1
$ log-format -json json
Log format is not reconfigurable on this logger
exit status 1
$ log-format -pretty json
Log format is not reconfigurable on this logger
exit status 1
$ log-format text
Log format is not reconfigurable on this logger
exit status 1
$ log-format -json text
Log format is not reconfigurable on this logger
exit status 1
$ log-format -pretty text
Log format is not reconfigurable on this logger
exit status 1
$ change-remote
No vpn address was provided
exit status 1
$ change-remote -json
flag provided but not defined: -json
Usage:
-address string
The new remote address, ip:port
exit status 1
$ change-remote -pretty
flag provided but not defined: -pretty
Usage:
-address string
The new remote address, ip:port
exit status 1
$ change-remote 10.222.0.9
No address was provided
exit status 1
$ change-remote -json 10.222.0.9
flag provided but not defined: -json
Usage:
-address string
The new remote address, ip:port
exit status 1
$ change-remote -pretty 10.222.0.9
flag provided but not defined: -pretty
Usage:
-address string
The new remote address, ip:port
exit status 1
$ change-remote -address 10.94.0.10:4242 10.222.0.9
Could not find tunnel for vpn address: 10.222.0.9
exit status 1
$ change-remote -json -address 10.94.0.10:4242 10.222.0.9
flag provided but not defined: -json
Usage:
-address string
The new remote address, ip:port
exit status 1
$ change-remote -pretty -address 10.94.0.10:4242 10.222.0.9
flag provided but not defined: -pretty
Usage:
-address string
The new remote address, ip:port
exit status 1
$ change-remote -address 10.94.0.10:4242 10.222.0.250
Could not find tunnel for vpn address: 10.222.0.250
exit status 1
$ change-remote -json -address 10.94.0.10:4242 10.222.0.250
flag provided but not defined: -json
Usage:
-address string
The new remote address, ip:port
exit status 1
$ change-remote -pretty -address 10.94.0.10:4242 10.222.0.250
flag provided but not defined: -pretty
Usage:
-address string
The new remote address, ip:port
exit status 1
$ create-tunnel
No vpn address was provided
exit status 1
$ create-tunnel -json
flag provided but not defined: -json
Usage:
-address string
Optionally provide a real remote address, ip:port
exit status 1
$ create-tunnel -pretty
flag provided but not defined: -pretty
Usage:
-address string
Optionally provide a real remote address, ip:port
exit status 1
$ create-tunnel 10.222.0.9
Tunnel already handshaking
exit status 1
$ create-tunnel -json 10.222.0.9
flag provided but not defined: -json
Usage:
-address string
Optionally provide a real remote address, ip:port
exit status 1
$ create-tunnel -pretty 10.222.0.9
flag provided but not defined: -pretty
Usage:
-address string
Optionally provide a real remote address, ip:port
exit status 1
$ create-tunnel -address 10.94.0.10:4242 10.222.0.10
Created
$ create-tunnel -json -address 10.94.0.10:4242 10.222.0.10
flag provided but not defined: -json
Usage:
-address string
Optionally provide a real remote address, ip:port
exit status 1
$ create-tunnel -pretty -address 10.94.0.10:4242 10.222.0.10
flag provided but not defined: -pretty
Usage:
-address string
Optionally provide a real remote address, ip:port
exit status 1
$ create-tunnel not-an-address
The provided vpn address could not be parsed: not-an-address
exit status 1
$ create-tunnel -json not-an-address
flag provided but not defined: -json
Usage:
-address string
Optionally provide a real remote address, ip:port
exit status 1
$ create-tunnel -pretty not-an-address
flag provided but not defined: -pretty
Usage:
-address string
Optionally provide a real remote address, ip:port
exit status 1
$ close-tunnel
No vpn address was provided
exit status 1
$ close-tunnel -json
flag provided but not defined: -json
Usage:
-local-only
Disables notifying the remote that the tunnel is shutting down
exit status 1
$ close-tunnel -pretty
flag provided but not defined: -pretty
Usage:
-local-only
Disables notifying the remote that the tunnel is shutting down
exit status 1
$ close-tunnel 10.222.0.250
Could not find tunnel for vpn address: 10.222.0.250
exit status 1
$ close-tunnel -json 10.222.0.250
flag provided but not defined: -json
Usage:
-local-only
Disables notifying the remote that the tunnel is shutting down
exit status 1
$ close-tunnel -pretty 10.222.0.250
flag provided but not defined: -pretty
Usage:
-local-only
Disables notifying the remote that the tunnel is shutting down
exit status 1
$ close-tunnel not-an-address
The provided vpn address could not be parsed: not-an-address
exit status 1
$ close-tunnel -json not-an-address
flag provided but not defined: -json
Usage:
-local-only
Disables notifying the remote that the tunnel is shutting down
exit status 1
$ close-tunnel -pretty not-an-address
flag provided but not defined: -pretty
Usage:
-local-only
Disables notifying the remote that the tunnel is shutting down
exit status 1
$ help
Available commands:
change-remote - Changes the remote address used in the tunnel for the provided vpn addr
close-tunnel - Closes a tunnel for the provided vpn addr
create-tunnel - Creates a tunnel for the provided vpn address
device-info - Prints information about the network device.
help - prints available commands or help <command> for specific usage info
list-hostmap - List all known previously connected hosts
list-lighthouse-addrmap - List all lighthouse map entries
list-pending-hostmap - List all handshaking hosts
log-format - Gets or sets the current log format
log-level - Gets or sets the current log level
logout - Ends the current session
print-cert - Prints the current certificate being used or the certificate for the provided vpn addr
print-relays - Prints json details about all relay info
print-tunnel - Prints json details about a tunnel for the provided vpn addr
query-lighthouse - Query the lighthouses for the provided vpn address
reload - Reloads configuration from disk, same as sending HUP to the process
version - Prints the currently running version of nebula
$ help -json
Command not available -json
exit status 1
$ help -pretty
Command not available -pretty
exit status 1
$ help print-cert
print-cert - Prints the current certificate being used or the certificate for the provided vpn addr
-json
outputs as json
-pretty
pretty prints json, assumes -json
-raw
raw prints the PEM encoded certificate, not compatible with -json or -pretty
$ help -json print-cert
Command not available -json
exit status 1
$ help -pretty print-cert
Command not available -pretty
exit status 1
$ help logout
logout - Ends the current session
$ help -json logout
Command not available -json
exit status 1
$ help -pretty logout
Command not available -pretty
exit status 1
$ logout -h
logout - Ends the current session
$ logout -json -h
logout - Ends the current session
$ logout -pretty -h
logout - Ends the current session
$ not-a-command 10.222.0.9
did not understand: not-a-command
Available commands:
change-remote - Changes the remote address used in the tunnel for the provided vpn addr
close-tunnel - Closes a tunnel for the provided vpn addr
create-tunnel - Creates a tunnel for the provided vpn address
device-info - Prints information about the network device.
help - prints available commands or help <command> for specific usage info
list-hostmap - List all known previously connected hosts
list-lighthouse-addrmap - List all lighthouse map entries
list-pending-hostmap - List all handshaking hosts
log-format - Gets or sets the current log format
log-level - Gets or sets the current log level
logout - Ends the current session
print-cert - Prints the current certificate being used or the certificate for the provided vpn addr
print-relays - Prints json details about all relay info
print-tunnel - Prints json details about a tunnel for the provided vpn addr
query-lighthouse - Query the lighthouses for the provided vpn address
reload - Reloads configuration from disk, same as sending HUP to the process
version - Prints the currently running version of nebula
exit status 1
$ not-a-command -json 10.222.0.9
did not understand: not-a-command
Available commands:
change-remote - Changes the remote address used in the tunnel for the provided vpn addr
close-tunnel - Closes a tunnel for the provided vpn addr
create-tunnel - Creates a tunnel for the provided vpn address
device-info - Prints information about the network device.
help - prints available commands or help <command> for specific usage info
list-hostmap - List all known previously connected hosts
list-lighthouse-addrmap - List all lighthouse map entries
list-pending-hostmap - List all handshaking hosts
log-format - Gets or sets the current log format
log-level - Gets or sets the current log level
logout - Ends the current session
print-cert - Prints the current certificate being used or the certificate for the provided vpn addr
print-relays - Prints json details about all relay info
print-tunnel - Prints json details about a tunnel for the provided vpn addr
query-lighthouse - Query the lighthouses for the provided vpn address
reload - Reloads configuration from disk, same as sending HUP to the process
version - Prints the currently running version of nebula
exit status 1
$ not-a-command -pretty 10.222.0.9
did not understand: not-a-command
Available commands:
change-remote - Changes the remote address used in the tunnel for the provided vpn addr
close-tunnel - Closes a tunnel for the provided vpn addr
create-tunnel - Creates a tunnel for the provided vpn address
device-info - Prints information about the network device.
help - prints available commands or help <command> for specific usage info
list-hostmap - List all known previously connected hosts
list-lighthouse-addrmap - List all lighthouse map entries
list-pending-hostmap - List all handshaking hosts
log-format - Gets or sets the current log format
log-level - Gets or sets the current log level
logout - Ends the current session
print-cert - Prints the current certificate being used or the certificate for the provided vpn addr
print-relays - Prints json details about all relay info
print-tunnel - Prints json details about a tunnel for the provided vpn addr
query-lighthouse - Query the lighthouses for the provided vpn address
reload - Reloads configuration from disk, same as sending HUP to the process
version - Prints the currently running version of nebula
exit status 1
$ reload
Reloading config
Config reload failed, the new config did not load: no config files found at
exit status 1
$ reload -json
Reloading config
Config reload failed, the new config did not load: no config files found at
exit status 1
$ reload -pretty
Reloading config
Config reload failed, the new config did not load: no config files found at
exit status 1
+2 -2
View File
@@ -129,7 +129,7 @@ func TestCertUpgrade(t *testing.T) {
}
r.Logf("reload new v2-only config")
err = myC.ReloadConfigString(string(cb))
err = myC.ReloadConfigString(string(cb)).Err
assert.NoError(t, err)
r.Log("yay, spin until their sees it")
waitStart := time.Now()
@@ -222,7 +222,7 @@ func TestCertDowngrade(t *testing.T) {
}
r.Logf("reload new v1-only config")
err = myC.ReloadConfigString(string(cb))
err = myC.ReloadConfigString(string(cb)).Err
assert.NoError(t, err)
r.Log("yay, spin until their sees it")
waitStart := time.Now()
+26 -19
View File
@@ -8,6 +8,7 @@ pki:
cert: /etc/nebula/host.crt
key: /etc/nebula/host.key
# blocklist is a list of certificate fingerprints that we will refuse to talk to
# Entries are the hex fingerprints shown by `nebula-cert print`, an invalid entry fails startup and reload.
#blocklist:
# - c99d4e650533b92061b09918e838a5a0a6aaee21eed1d12fd937682865936c72
# disconnect_invalid is a toggle to force a client to be disconnected if the certificate is expired or invalid.
@@ -43,7 +44,7 @@ static_host_map:
lighthouse:
# am_lighthouse is used to enable lighthouse functionality for a node. This should ONLY be true on nodes
# you have configured to be lighthouses in your network
# you have configured to be lighthouses in your network. Not reloadable.
am_lighthouse: false
# serve_dns optionally starts a dns listener that responds to various queries and can even be
# delegated to for resolution
@@ -126,8 +127,9 @@ lighthouse:
# Port Nebula will be listening on. The default here is 4242. For a lighthouse node, the port should be defined,
# however using port 0 will dynamically assign a port and is recommended for roaming nodes.
listen:
# To listen on only ipv4, use "0.0.0.0"
# To listen on only ipv4, use "0.0.0.0". Not reloadable.
host: "::"
# Not reloadable.
port: 4242
# Sets the max number of packets to pull from the kernel for each syscall (under systems that support recvmmsg)
# default is 64, does not support reload
@@ -135,12 +137,13 @@ listen:
# superpacket, so the receive scratch is batch * 64KiB per listening socket (~4MiB per routine at the
# default of 64). Lower this to trade peak per-syscall throughput for memory on constrained hosts.
#batch: 64
# Linux only. Reloadable, but removing a value leaves the socket at the old size until Nebula restarts.
# Configure socket buffers for the udp side (outside), leave unset to use the system defaults. Values will be doubled by the kernel
# Default is net.core.rmem_default and net.core.wmem_default (/proc/sys/net/core/rmem_default and /proc/sys/net/core/rmem_default)
# Maximum is limited by memory in the system, SO_RCVBUFFORCE and SO_SNDBUFFORCE is used to avoid having to raise the system wide
# max, net.core.rmem_max and net.core.wmem_max
#read_buffer: 10485760
#write_buffer: 10485760
read_buffer: 3412512
write_buffer: 3412512
# On Windows only
# When true, Nebula installs a WFP (Windows Filtering Platform) PERMIT filter scoped to UDP at the listener port.
@@ -179,7 +182,7 @@ listen:
# Currently, this defaults to 1 which means we have 1 tun queue reader and 1
# UDP queue reader. Setting this above one will set IFF_MULTI_QUEUE on the tun
# device and SO_REUSEPORT on the UDP socket to allow multiple queues.
# This option is only supported on Linux.
# This option is only supported on Linux. Not reloadable.
#routines: 1
punchy:
@@ -203,6 +206,7 @@ punchy:
# Cipher allows you to choose between the available ciphers for your network. Options are chachapoly or aes
# IMPORTANT: this value must be identical on ALL NODES/LIGHTHOUSES. We do not/will not support use of different ciphers simultaneously!
# Not reloadable.
#cipher: aes
# Preferred ranges is used to define a hint about the local network ranges, which speeds up discovering the fastest
@@ -229,12 +233,6 @@ punchy:
# Trusted SSH CA public keys. These are the public keys of the CAs that are allowed to sign SSH keys for access.
#trusted_cas:
#- "ssh public key string"
# sandbox_dir restricts file paths for profiling commands (start-cpu-profile, save-heap-profile,
# save-mutex-profile) to the specified directory. Relative paths will be resolved within this directory,
# and absolute paths outside of it will be rejected. Default is $TMP/nebula-debug.
# The directory is NOT automatically created.
# Overriding this to "" is the same as "/" and will allow overwriting any path on the host.
#sandbox_dir: /var/tmp/nebula-debug
# EXPERIMENTAL: relay support for networks that can't establish direct connections.
relay:
@@ -253,18 +251,25 @@ relay:
# Configure the private interface. Note: addr is baked into the nebula certificate
tun:
# When tun is disabled, a lighthouse can be started without a local tun interface (and therefore without root)
# Not reloadable.
disabled: false
# Name of the device. If not set, a default will be chosen by the OS.
# For macOS: if set, must be in the form `utun[0-9]+`.
# For NetBSD: Required to be set, must be in the form `tun[0-9]+`
# For OpenBSD and NetBSD: if set, must be in the form `tun[0-9]+`, otherwise the first free one is used.
# Not reloadable.
dev: nebula1
# Toggles forwarding of local broadcast packets, the address of which depends on the ip/mask encoded in pki.cert
# Not reloadable.
drop_local_broadcast: false
# Toggles forwarding of multicast packets
# Toggles forwarding of multicast packets. Not reloadable.
drop_multicast: false
# Sets the transmit queue length, if you notice lots of transmit drops on the tun it may help to raise this number. Default is 500
tx_queue: 500
# Sets the transmit queue length, if you notice lots of transmit drops on the tun it may help to raise this number. Default is 5000
# Not reloadable.
tx_queue: 5000
# Default MTU for every packet, safe setting is (and the default) 1300 for internet based traffic
# Anything above 8937, here or on a route below, is capped to 8937 with a warning
# Reloadable on Linux only. Where the tun is handed to nebula by a host app (iOS, Android, macOS network extensions)
# the host app owns the MTU.
mtu: 1300
# the use_offloads setting controls if Nebula will attempt to enable GSO and GRO for the tun device. Linux only, not reloadable.
@@ -357,7 +362,7 @@ tun:
#use_system_route_table: false
# Buffer size for reading routes updates. 0 means default system buffer size. (/proc/sys/net/core/rmem_default).
# If using massive routes updates, for example BGP, you may need to increase this value to avoid packet loss.
# SO_RCVBUFFORCE is used to avoid having to raise the system wide max
# SO_RCVBUFFORCE is used to avoid having to raise the system wide max. Not reloadable.
#use_system_route_table_buffer_size: 0
# Configure logging level
@@ -407,14 +412,16 @@ logging:
# Handshakes are sent to all known addresses at each interval with a linear backoff,
# Wait try_interval after the 1st attempt, 2 * try_interval after the 2nd, etc, until the handshake is older than timeout
# A 100ms interval with the default 10 retries will give a handshake 5.5 seconds to resolve before timing out
# Not reloadable.
#try_interval: 100ms
#retries: 20
# query_buffer is the size of the buffer channel for querying lighthouses
#query_buffer: 64
# query_buffer is the size of the buffer channel for querying lighthouses, drops are recorded in the lighthouse.query.dropped metric.
# Not reloadable.
#query_buffer: 1024
# trigger_buffer is the size of the buffer channel for quickly sending handshakes
# after receiving the response for lighthouse queries
# after receiving the response for lighthouse queries. Not reloadable.
#trigger_buffer: 64
# Tunnel manager settings
+4 -4
View File
@@ -1,13 +1,12 @@
module github.com/slackhq/nebula
go 1.26.0
go 1.27.2
require (
dario.cat/mergo v1.0.2
filippo.io/bigmod v0.1.0
github.com/DefinedNet/netlink v1.3.2-dn.1
github.com/anmitsu/go-shlex v0.0.0-20200514113438-38f4b401e2be
github.com/armon/go-radix v1.0.0
github.com/flynn/noise v1.1.0
github.com/gaissmai/bart v0.29.0
github.com/gogo/protobuf v1.3.2
@@ -33,7 +32,7 @@ require (
golang.zx2c4.com/wireguard/windows v1.0.1
google.golang.org/protobuf v1.36.12
gopkg.in/yaml.v3 v3.0.1
gvisor.dev/gvisor v0.0.0-20240423190808-9d7a357edefe
gvisor.dev/gvisor v0.0.0-20261014193016-04d31b385980
)
require (
@@ -45,5 +44,6 @@ require (
github.com/prometheus/common v0.70.1 // indirect
github.com/prometheus/procfs v0.21.1 // indirect
github.com/vishvananda/netns v0.0.5 // indirect
golang.org/x/time v0.5.0 // indirect
golang.org/x/exp v0.0.0-20250711185948-6ae5c78190dc // indirect
golang.org/x/time v0.15.0 // indirect
)
+6 -6
View File
@@ -12,8 +12,6 @@ github.com/alecthomas/units v0.0.0-20190717042225-c3de453c63f4/go.mod h1:ybxpYRF
github.com/alecthomas/units v0.0.0-20190924025748-f65c72e2690d/go.mod h1:rBZYJk541a8SKzHPHnH3zbiI+7dagKZ0cgpgrD7Fyho=
github.com/anmitsu/go-shlex v0.0.0-20200514113438-38f4b401e2be h1:9AeTilPcZAjCFIImctFaOjnTIavg87rW78vTPkQqLI8=
github.com/anmitsu/go-shlex v0.0.0-20200514113438-38f4b401e2be/go.mod h1:ySMOLuWl6zY27l47sB3qLNK6tF2fkHG55UZxx8oIVo4=
github.com/armon/go-radix v1.0.0 h1:F4z6KzEeeQIMeLFa97iZU6vupzoecKdU5TX24SNppXI=
github.com/armon/go-radix v1.0.0/go.mod h1:ufUuZ+zHj4x4TnLV4JWEpy2hxWSpsRywHrMgIH9cCH8=
github.com/beorn7/perks v0.0.0-20180321164747-3a771d992973/go.mod h1:Dwedo/Wpr24TaqPxmxbtue+5NUziq4I4S80YR8gNf3Q=
github.com/beorn7/perks v1.0.0/go.mod h1:KWe93zE9D1o94FZ5RNwFwVgaQK1VOXiVxmqh+CedLV8=
github.com/beorn7/perks v1.0.1 h1:VlbKKnNfV8bJzeqoa4cOKqO6bYr3WgKZxO8Z16+hsOM=
@@ -157,6 +155,8 @@ golang.org/x/crypto v0.0.0-20200622213623-75b288015ac9/go.mod h1:LzIPMQfyMNhhGPh
golang.org/x/crypto v0.0.0-20210322153248-0c34fe9e7dc2/go.mod h1:T9bdIzuCu7OtxOm1hfPfRQxPLYneinmdGuTeoZ9dtd4=
golang.org/x/crypto v0.57.0 h1:3ZVCjf8Ggz7zneR/EHRVx68Ctf+2pmIMP2UFhh9cC6M=
golang.org/x/crypto v0.57.0/go.mod h1:Fdz0i5U6CoizGwLda9DttjSk6qlZo25zYNtR+ycvuZA=
golang.org/x/exp v0.0.0-20250711185948-6ae5c78190dc h1:TS73t7x3KarrNd5qAipmspBDS1rkMcgVG/fS1aRb4Rc=
golang.org/x/exp v0.0.0-20250711185948-6ae5c78190dc/go.mod h1:A+z0yzpGtvnG90cToK5n2tu8UJVP2XUATh+r+sfOOOc=
golang.org/x/lint v0.0.0-20200302205851-738671d3881b/go.mod h1:3xt1FjdF8hUf6vQPIChWIBhFzV8gjjsPE/fR3IyQdNY=
golang.org/x/mod v0.1.1-0.20191105210325-c90efee705ee/go.mod h1:QqPTAvyqsEbceGzBzNggFXnrqF1CaUcvgkdR5Ot7KZg=
golang.org/x/mod v0.2.0/go.mod h1:s0Qsj1ACt9ePp/hMypM3fl4fZqREWJwdYDEqhRiZZUA=
@@ -203,8 +203,8 @@ golang.org/x/term v0.46.0/go.mod h1:+K02xbkittuwc0Am4abfA3Fc+XRGXkvBXNO88NCXPoc=
golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
golang.org/x/text v0.3.2/go.mod h1:bEr9sfX3Q8Zfm5fL9x+3itogRgK3+ptLWKqgva+5dAk=
golang.org/x/text v0.3.3/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
golang.org/x/time v0.5.0 h1:o7cqy6amK/52YcAKIPlM3a+Fpj35zvRj2TP+e1xFSfk=
golang.org/x/time v0.5.0/go.mod h1:3BpzKBy/shNhVucY/MWOyx10tF3SFh9QdLuxbVysPQM=
golang.org/x/time v0.15.0 h1:bbrp8t3bGUeFOx08pvsMYRTCVSMk89u4tKbNOZbp88U=
golang.org/x/time v0.15.0/go.mod h1:Y4YMaQmXwGQZoFaVFk4YpCt4FLQMYKZe9oeV/f4MSno=
golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
golang.org/x/tools v0.0.0-20191119224855-298f0cb1881e/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
golang.org/x/tools v0.0.0-20200130002326-2f3ba24bd6e7/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
@@ -243,5 +243,5 @@ gopkg.in/yaml.v2 v2.3.0/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA=
gopkg.in/yaml.v3 v3.0.1/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
gvisor.dev/gvisor v0.0.0-20240423190808-9d7a357edefe h1:fre4i6mv4iBuz5lCMOzHD1rH1ljqHWSICFmZRbbgp3g=
gvisor.dev/gvisor v0.0.0-20240423190808-9d7a357edefe/go.mod h1:sxc3Uvk/vHcd3tj7/DHVBoR5wvWT/MmRq2pj7HRJnwU=
gvisor.dev/gvisor v0.0.0-20261014193016-04d31b385980 h1:ribyrTYQ/QYeUIs8x8bVLYwGyepa7RnHLqJzegT5YYk=
gvisor.dev/gvisor v0.0.0-20261014193016-04d31b385980/go.mod h1:pK/dapF5Bf3KGW7/kd5CI5IC4lYh9FMFY1qETAIGLXk=
+18 -5
View File
@@ -31,6 +31,7 @@ package nebula
import (
"bufio"
"context"
"fmt"
"net"
"strconv"
@@ -51,17 +52,29 @@ type graphiteConfigExport struct {
Percentiles []float64 // Percentiles to export from timers and histograms
}
// graphiteTimeout bounds one submission, the connect and the write together. An export is a few KB, a host that can't
// take that in this long isn't going to.
const graphiteTimeout = 30 * time.Second
// graphiteOnce performs a single submission to Graphite, returning a non-nil
// error on failed connections.
func graphiteOnce(c graphiteConfigExport) error {
// error if it could not connect or write all of it within timeout. It gives up
// as soon as ctx is done.
func graphiteOnce(ctx context.Context, c graphiteConfigExport, timeout time.Duration) error {
now := time.Now().Unix()
du := float64(c.DurationUnit)
flushSeconds := float64(c.FlushInterval) / float64(time.Second)
conn, err := net.DialTCP("tcp", nil, c.Addr)
deadline := time.Now().Add(timeout)
d := net.Dialer{Deadline: deadline}
conn, err := d.DialContext(ctx, "tcp", c.Addr.String())
if err != nil {
return err
}
defer conn.Close()
if err := conn.SetDeadline(deadline); err != nil {
return err
}
stop := context.AfterFunc(ctx, func() { _ = conn.Close() })
defer stop()
w := bufio.NewWriter(conn)
c.Registry.Each(func(name string, i any) {
switch metric := i.(type) {
@@ -111,7 +124,7 @@ func graphiteOnce(c graphiteConfigExport) error {
fmt.Fprintf(w, "%s.%s.fifteen-minute %.2f %d\n", c.Prefix, name, t.Rate15(), now)
fmt.Fprintf(w, "%s.%s.mean-rate %.2f %d\n", c.Prefix, name, t.RateMean(), now)
}
w.Flush()
})
return nil
// A failed write sticks, so this reports the first one
return w.Flush()
}
+315
View File
@@ -0,0 +1,315 @@
package nebula
import (
"bytes"
"context"
"errors"
"fmt"
"io"
"net"
"strings"
"sync"
"testing"
"time"
"github.com/rcrowley/go-metrics"
"github.com/slackhq/nebula/test"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
// graphiteHost accepts connections and hands each one to the test
type graphiteHost struct {
ln net.Listener
conns chan net.Conn
mu sync.Mutex
closed bool
open []net.Conn
}
// newGraphiteHost accepts connections for the test. A non-zero window shrinks each connection's receive buffer, so a
// sender that is not read from blocks after a little whatever the host's TCP tuning. Only for tests that need a stall,
// a small window makes reading very slow on Linux.
func newGraphiteHost(t *testing.T, window int) *graphiteHost {
t.Helper()
ln, err := net.Listen("tcp", "127.0.0.1:0")
require.NoError(t, err)
h := &graphiteHost{ln: ln, conns: make(chan net.Conn, 8)}
done := make(chan struct{})
t.Cleanup(func() {
_ = ln.Close()
h.mu.Lock()
h.closed = true
for _, c := range h.open {
_ = c.Close()
}
h.mu.Unlock()
close(done)
})
go func() {
for {
c, err := ln.Accept()
if err != nil {
return
}
h.mu.Lock()
if h.closed {
h.mu.Unlock()
_ = c.Close()
return
}
h.open = append(h.open, c)
h.mu.Unlock()
if tc, ok := c.(*net.TCPConn); ok && window > 0 {
_ = tc.SetReadBuffer(window)
}
select {
case h.conns <- c:
case <-done:
return
}
}
}()
return h
}
func (h *graphiteHost) addr(t *testing.T) *net.TCPAddr {
a, err := net.ResolveTCPAddr("tcp", h.ln.Addr().String())
require.NoError(t, err)
return a
}
// readExport reads what one send wrote, bounded so a send that never finishes fails the test rather than hanging it
func readExport(t *testing.T, c net.Conn) []byte {
t.Helper()
require.NoError(t, c.SetReadDeadline(time.Now().Add(10*time.Second)))
b, err := io.ReadAll(c)
require.NoError(t, err)
return b
}
func (h *graphiteHost) next(t *testing.T, wait time.Duration) net.Conn {
t.Helper()
select {
case c := <-h.conns:
return c
case <-time.After(wait):
t.Fatal("no connection")
return nil
}
}
// lockedBuffer is a log destination a test can read while the code under test writes to it from another goroutine
type lockedBuffer struct {
mu sync.Mutex
buf bytes.Buffer
}
func (b *lockedBuffer) Write(p []byte) (int, error) {
b.mu.Lock()
defer b.mu.Unlock()
return b.buf.Write(p)
}
func (b *lockedBuffer) String() string {
b.mu.Lock()
defer b.mu.Unlock()
return b.buf.String()
}
func testGraphiteConfig(addr *net.TCPAddr, r metrics.Registry) graphiteConfigExport {
return graphiteConfigExport{Addr: addr, Registry: r, FlushInterval: time.Second, DurationUnit: time.Nanosecond, Prefix: "nebula"}
}
// Enough metrics that one export overflows the socket buffers on any platform, so a send to a host that stops
// reading blocks
func bigGraphiteRegistry(r metrics.Registry, t *testing.T) {
for i := range 200000 {
name := fmt.Sprintf("big.gauge.%d", i)
metrics.GetOrRegisterGauge(name, r).Update(int64(i))
t.Cleanup(func() { r.Unregister(name) })
}
}
func TestGraphiteOnce_sends(t *testing.T) {
r := metrics.NewRegistry()
metrics.GetOrRegisterGauge("g", r).Update(7)
metrics.GetOrRegisterCounter("c", r).Inc(3)
h := newGraphiteHost(t, 0)
require.NoError(t, graphiteOnce(t.Context(), testGraphiteConfig(h.addr(t), r), time.Minute))
b := readExport(t, h.next(t, 5*time.Second))
assert.Contains(t, string(b), "nebula.g.value 7 ")
assert.Contains(t, string(b), "nebula.c.count 3 ")
assert.Contains(t, string(b), "nebula.c.count_ps 3.00 ")
}
// A host that accepts and never reads is given up on, and the failed write is reported
func TestGraphiteOnce_stalledHostTimesOut(t *testing.T) {
r := metrics.NewRegistry()
bigGraphiteRegistry(r, t)
h := newGraphiteHost(t, 4096)
done := make(chan error, 1)
cfg := testGraphiteConfig(h.addr(t), r)
go func() { done <- graphiteOnce(t.Context(), cfg, 300*time.Millisecond) }()
h.next(t, 5*time.Second) // accepted, never read
select {
case err := <-done:
require.Error(t, err)
case <-time.After(10 * time.Second):
t.Fatal("the send did not give up on a host that stopped reading")
}
}
// The timeout covers the whole send, a host that keeps reading but too slowly is still given up on
func TestGraphiteOnce_timeoutCoversTheWholeSend(t *testing.T) {
r := metrics.NewRegistry()
bigGraphiteRegistry(r, t)
h := newGraphiteHost(t, 0)
done := make(chan error, 1)
start := time.Now()
cfg := testGraphiteConfig(h.addr(t), r)
go func() { done <- graphiteOnce(t.Context(), cfg, 500*time.Millisecond) }()
c := h.next(t, 5*time.Second)
stopReading := make(chan struct{})
defer close(stopReading)
// 32KB every 10ms keeps every write well inside the timeout, the whole export takes seconds
go func() {
buf := make([]byte, 32<<10)
for {
select {
case <-stopReading:
return
case <-time.After(10 * time.Millisecond):
}
if _, err := c.Read(buf); err != nil {
return
}
}
}()
select {
case err := <-done:
require.Error(t, err)
assert.Less(t, time.Since(start), 5*time.Second)
case <-time.After(10 * time.Second):
t.Fatal("a host that kept reading slowly held the send past its timeout")
}
}
func TestGraphiteOnce_cancelledCtxDoesNotConnect(t *testing.T) {
h := newGraphiteHost(t, 0)
ctx, cancel := context.WithCancel(t.Context())
cancel()
require.Error(t, graphiteOnce(ctx, testGraphiteConfig(h.addr(t), metrics.NewRegistry()), time.Minute))
select {
case <-h.conns:
t.Fatal("connected on a cancelled ctx")
case <-time.After(200 * time.Millisecond):
}
}
// A stop or reload abandons a stuck send rather than waiting out its timeout
func TestGraphiteOnce_cancelAbandonsSend(t *testing.T) {
r := metrics.NewRegistry()
bigGraphiteRegistry(r, t)
h := newGraphiteHost(t, 4096)
ctx, cancel := context.WithCancel(t.Context())
done := make(chan error, 1)
cfg := testGraphiteConfig(h.addr(t), r)
go func() { done <- graphiteOnce(ctx, cfg, time.Minute) }()
h.next(t, 5*time.Second)
select {
case err := <-done:
t.Fatalf("the send should be stuck on a host that is not reading, it ended with %v", err)
case <-time.After(200 * time.Millisecond):
}
cancel()
select {
case err := <-done:
require.Error(t, err)
case <-time.After(5 * time.Second):
t.Fatal("the send is still stuck after its ctx was cancelled")
}
}
// buildRuntime hands graphite the configured prefix and interval
func TestStatsServer_graphiteExportUsesConfig(t *testing.T) {
const name = "graphite.wiring.test"
metrics.GetOrRegisterCounter(name, nil).Inc(8)
t.Cleanup(func() { metrics.DefaultRegistry.Unregister(name) })
h := newGraphiteHost(t, 0)
s, c := newTestStatsServer(t)
setStatsConfig(c, map[string]any{"type": "graphite", "interval": "4s", "prefix": "pfx", "host": h.ln.Addr().String()})
cfg, err := loadStatsConfig(c)
require.NoError(t, err)
s.runMu.Lock()
fns, _ := s.buildRuntime(t.Context(), cfg)
s.runMu.Unlock()
for _, fn := range fns {
fn()
}
b := readExport(t, h.next(t, 5*time.Second))
assert.Contains(t, string(b), "pfx."+name+".count 8 ")
assert.Contains(t, string(b), "pfx."+name+".count_ps 2.00 ")
}
func TestStatsServer_graphiteFailureIsLogged(t *testing.T) {
logs := &lockedBuffer{}
s, c := newTestStatsServer(t)
s.l = test.NewLoggerWithOutput(logs)
setStatsConfig(c, map[string]any{"type": "graphite", "interval": "1h", "host": "127.0.0.1:" + freeTCPPort(t)})
cfg, err := loadStatsConfig(c)
require.NoError(t, err)
s.runMu.Lock()
fns, _ := s.buildRuntime(t.Context(), cfg)
s.runMu.Unlock()
for _, fn := range fns {
fn()
}
assert.Contains(t, logs.String(), "Graphite export failed")
}
// Stop abandons the running export, so it goes through Start rather than handing buildRuntime a ctx, and a stop is
// not logged as a failure
func TestStatsServer_stopAbandonsGraphiteSend(t *testing.T) {
bigGraphiteRegistry(metrics.DefaultRegistry, t)
h := newGraphiteHost(t, 4096)
logs := &lockedBuffer{}
s, c := newTestStatsServer(t)
s.l = test.NewLoggerWithOutput(logs)
setStatsConfig(c, map[string]any{"type": "graphite", "interval": "50ms", "host": h.ln.Addr().String()})
require.NoError(t, s.reload(c, true))
started := make(chan struct{})
go func() {
s.Start()
close(started)
}()
stuck := h.next(t, 30*time.Second)
s.Stop()
select {
case <-started:
case <-time.After(5 * time.Second):
t.Fatal("Start did not return after Stop")
}
// A closed sender resets the connection once it sees this, one still sending just buffers it. Either way the
// read below doesn't have to drain the export through the small window to find out
_, err := stuck.Write([]byte("x"))
require.NoError(t, err)
require.NoError(t, stuck.SetReadDeadline(time.Now().Add(10*time.Second)))
n, err := io.Copy(io.Discard, stuck)
var ne net.Error
if errors.As(err, &ne) && ne.Timeout() {
t.Fatalf("the send kept the connection after Stop, read %d bytes", n)
}
assert.Less(t, n, int64(1<<20), "the send kept going after Stop")
// The capture loop logs on its own goroutine after Start has returned, give it the chance
assert.Never(t, func() bool { return strings.Contains(logs.String(), "Graphite export failed") }, 500*time.Millisecond, 10*time.Millisecond, "a stop is not a failure")
}
+8 -11
View File
@@ -54,7 +54,8 @@ type HandshakeConfig struct {
}
type HandshakeManager struct {
// Mutex for interacting with the vpnIps and indexes maps
// Mutex for interacting with the vpnIps and indexes maps.
// Lock order is hostmap, then this lock, then the lighthouse, then a RemoteList.
sync.RWMutex
vpnIps map[netip.Addr]*HandshakeHostInfo
@@ -259,14 +260,6 @@ func (hm *HandshakeManager) handleOutbound(vpnIp netip.Addr, lighthouseTriggered
"style": header.SubTypeName(header.Handshake, hh.machine.Subtype()),
}
// Get a remotes object if we don't already have one.
// This is mainly to protect us as this should never be the case
// NB ^ This comment doesn't jive. It's how the thing gets initialized.
// It's the common path. Should it update every time, in case a future LH query/queries give us more info?
if hostinfo.remotes == nil {
hostinfo.remotes = hm.lightHouse.QueryCache([]netip.Addr{vpnIp})
}
remotes := hostinfo.remotes.CopyAddrs(hm.mainHostMap.GetPreferredRanges())
remotesHaveChanged := !slices.Equal(remotes, hh.lastRemotes)
@@ -371,6 +364,9 @@ func (hm *HandshakeManager) StartHandshake(vpnAddr netip.Addr, cacheCb func(*Han
hostinfo := &HostInfo{
vpnAddrs: []netip.Addr{vpnAddr},
HandshakePacket: make(map[uint8][]byte, 0),
// Attached here so a pending hostinfo never exists without one: SetRemote dereferences
// it, and create-tunnel -address can reach SetRemote before the handshake loop runs.
remotes: hm.lightHouse.QueryCache([]netip.Addr{vpnAddr}),
relayState: RelayState{
relays: nil,
relayForByAddr: map[netip.Addr]*Relay{},
@@ -946,13 +942,14 @@ func (hm *HandshakeManager) continueHandshake(via ViaSender, hh *HandshakeHostIn
hm.DeleteHostInfo(hostinfo)
hm.StartHandshake(hostinfo.vpnAddrs[0], func(newHH *HandshakeHostInfo) {
newHH.hostinfo.remotes = hostinfo.remotes
newHH.hostinfo.remotes.BlockRemote(via)
newHH.packetStore = hh.packetStore
hh.packetStore = []*cachedPacket{}
})
// sendCloseTunnel may take the hostmap lock, so it must not run under the handshake
// manager lock that StartHandshake holds around the callback.
hostinfo.vpnAddrs = vpnAddrs
f.sendCloseTunnel(hostinfo)
})
return
}
+68 -2
View File
@@ -11,6 +11,7 @@ import (
"github.com/slackhq/nebula/test"
"github.com/slackhq/nebula/udp"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
func Test_NewHandshakeManagerVpnIp(t *testing.T) {
@@ -42,8 +43,6 @@ func Test_NewHandshakeManagerVpnIp(t *testing.T) {
i2 := blah.StartHandshake(ip, nil)
assert.Same(t, i, i2)
i.remotes = NewRemoteList([]netip.Addr{}, nil)
// Adding something to pending should not affect the main hostmap
assert.Empty(t, mainHM.Hosts)
@@ -235,3 +234,70 @@ func TestHandleIncomingDispatch(t *testing.T) {
assert.Empty(t, hm.indexes, "orphan stage-2 must not create state")
})
}
// Test_PendingHandshakeHasRemotes covers create-tunnel -address: a pending hostinfo must already
// carry its RemoteList when StartHandshake runs its callback, so SetRemote there, before the
// handshake loop has run, is safe and the address becomes one the handshake is sent to.
func Test_PendingHandshakeHasRemotes(t *testing.T) {
l := test.NewLogger()
mainHM := newHostMap(l)
preferredRanges := []netip.Prefix{}
mainHM.preferredRanges.Store(&preferredRanges)
hm := NewHandshakeManager(l, mainHM, newTestLighthouse(), &udp.NoopConn{}, defaultHandshakeConfig)
vpnIp := netip.MustParseAddr("172.1.1.2")
remote := netip.MustParseAddrPort("192.168.1.9:4242")
hi := hm.StartHandshake(vpnIp, func(hh *HandshakeHostInfo) {
require.NotNil(t, hh.hostinfo.remotes, "a pending handshake must never be without its RemoteList")
hh.hostinfo.SetRemote(remote)
})
assert.Contains(t, hi.remotes.CopyAddrs(preferredRanges), remote, "the handshake must be sent to the supplied remote")
assert.Equal(t, remote, hi.GetRemote())
// A RemoteList keeps one learned address per host, so a second remote replaces the
// first — the same behaviour change-remote has on an established tunnel.
other := netip.MustParseAddrPort("192.168.1.10:4242")
hi.SetRemote(other)
assert.Equal(t, []netip.AddrPort{other}, hi.remotes.CopyAddrs(preferredRanges))
assert.Equal(t, other, hi.GetRemote())
}
// newRemotesRaceFixture builds a handshake manager with an established tunnel to vpnAddr whose
// RemoteList is the lighthouse's cached list, the state every teardown-during-handshake race starts from.
func newRemotesRaceFixture(t *testing.T, vpnAddr netip.Addr) (*HandshakeManager, *Interface, *HostInfo) {
t.Helper()
l := test.NewLogger()
mainHM := newHostMap(l)
preferredRanges := []netip.Prefix{}
mainHM.preferredRanges.Store(&preferredRanges)
lh := newTestLighthouse()
hm := NewHandshakeManager(l, mainHM, lh, &udp.NoopConn{}, defaultHandshakeConfig)
f := &Interface{hostMap: mainHM, lightHouse: lh, handshakeManager: hm, l: l}
hm.f = f
established := &HostInfo{
vpnAddrs: []netip.Addr{vpnAddr},
localIndexId: 1,
remoteIndexId: 101,
remotes: lh.QueryCache([]netip.Addr{vpnAddr}),
HandshakePacket: map[uint8][]byte{},
}
mainHM.Lock()
mainHM.unlockedAddHostInfo(established, f)
mainHM.Unlock()
return hm, f, established
}
// With no handshake pending, tearing down the last tunnel still evicts the list.
func Test_CloseTunnelEvictsWithoutPendingHandshake(t *testing.T) {
vpnAddr := netip.MustParseAddr("172.1.1.2")
_, f, established := newRemotesRaceFixture(t, vpnAddr)
f.closeTunnel(established)
assert.NotContains(t, f.hostMap.Hosts, vpnAddr)
assert.NotContains(t, f.lightHouse.addrMap, vpnAddr)
}
+5
View File
@@ -24,6 +24,11 @@ type m = map[string]any
const (
Version uint8 = 1
Len = 16
// MaxOverhead is the most a payload grows on its way to the underlay.
// A relayed packet has two headers and two AEAD tags, one for the inner tunnel and another for the relay tunnel.
// Every supported cipher (today) has a 16 byte tag.
MaxOverhead = 2 * (Len + 16)
)
type MessageType uint8
+7 -5
View File
@@ -316,9 +316,9 @@ type cachedPacketMetrics struct {
func NewHostMapFromConfig(l *slog.Logger, c *config.C) *HostMap {
hm := newHostMap(l)
hm.reload(c, true)
c.RegisterReloadCallback(func(c *config.C) {
hm.reload(c, false)
hm.reload(c, true, nil)
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
hm.reload(c, false, r)
})
l.Info("Main HostMap created", "preferredRanges", hm.GetPreferredRanges())
@@ -337,7 +337,8 @@ func newHostMap(l *slog.Logger) *HostMap {
}
}
func (hm *HostMap) reload(c *config.C, initial bool) {
// r is nil on the initial load
func (hm *HostMap) reload(c *config.C, initial bool, r *config.Reload) {
if initial || c.HasChanged("preferred_ranges") {
var preferredRanges []netip.Prefix
rawPreferredRanges := c.GetStringSlice("preferred_ranges", []string{})
@@ -348,8 +349,9 @@ func (hm *HostMap) reload(c *config.C, initial bool) {
if err != nil {
hm.l.Warn("Failed to parse preferred ranges, ignoring",
"error", err,
"range", rawPreferredRanges,
"range", rawPreferredRange,
)
r.Failed()
continue
}
+13
View File
@@ -447,3 +447,16 @@ func TestHostInfo_markOutOnly(t *testing.T) {
assert.True(t, h.markOut(6), "a relayed send must not consume the epoch edge")
assert.False(t, h.markOut(6))
}
// A preferred range that doesn't parse is skipped and the rest are used, the reload says one was skipped
func TestHostMap_reloadBadPreferredRangeFails(t *testing.T) {
l := test.NewLogger()
c := config.NewC(l)
require.NoError(t, c.LoadString("preferred_ranges: [10.0.0.0/8]\n"))
hm := NewHostMapFromConfig(l, c)
res := c.ReloadConfigString("preferred_ranges: [10.0.0.0/8, nope]\n")
require.NoError(t, res.Err)
assert.True(t, res.Failed)
assert.Equal(t, []netip.Prefix{netip.MustParsePrefix("10.0.0.0/8")}, hm.GetPreferredRanges())
}
+28 -11
View File
@@ -10,6 +10,7 @@ import (
"github.com/slackhq/nebula/firewall"
"github.com/slackhq/nebula/header"
"github.com/slackhq/nebula/iputil"
"github.com/slackhq/nebula/logging"
"github.com/slackhq/nebula/noiseutil"
"github.com/slackhq/nebula/overlay/batch"
"github.com/slackhq/nebula/overlay/tio"
@@ -198,8 +199,8 @@ func (f *Interface) sendInsideMessage(hostinfo *HostInfo, pkt tio.Packet, nb []b
}
err = tio.SegmentSuperpacket(pkt, func(seg []byte) error {
//relay header + header + plaintext + AEAD tag (16 bytes for both AES-GCM and ChaCha20-Poly1305) + relay tag
scratch := sendBatch.Reserve(header.Len + header.Len + len(seg) + 16 + 16)
//relay header + header + plaintext + AEAD tag + relay tag
scratch := sendBatch.Reserve(len(seg) + header.MaxOverhead)
innerPacket := f.sendInsideEncrypt(hostinfo, ci, seg, scratch[header.Len:], nb)
if innerPacket == nil {
@@ -517,7 +518,9 @@ func (f *Interface) SendVia(via *HostInfo, relay *Relay, ad, nb, out []byte, noc
err = f.writers[q].WriteTo(toSend, via.GetRemote())
if err != nil {
via.logger(f.l).Info("Failed to WriteTo in sendVia", "error", err)
if l, ok := logging.RateLimited(f.l, f.udpRelayTxErrLimiter); ok {
via.logger(l).Info("Failed to WriteTo in sendVia", "error", err)
}
}
}
@@ -526,16 +529,26 @@ func (f *Interface) sendNoMetrics(t header.MessageType, st header.MessageSubType
return
}
useRelay := !remote.IsValid() && !hostinfo.GetRemote().IsValid()
fullOut := out
// The encrypt appends to out in place and a relayed packet is then wrapped in place,
// so out has to fit the whole packet up front or the encrypt reallocates out from under SendVia.
// We cap the MTU of tun devices and routes we create ourselves, so only a host that set a bigger MTU than we allow gets here.
need := header.Len + len(p) + ci.eKey.Overhead()
if useRelay {
if len(out) < header.Len {
// out always has a capacity of mtu, but not always a length greater than the header.Len.
// Grow it to make sure the next operation works.
out = out[:header.Len]
// The relay tunnel runs the same cipher, and adds its own header and tag
need += header.Len + ci.eKey.Overhead()
}
if cap(out) < need {
if l, ok := logging.RateLimited(f.l, f.txOversizeLimiter); ok {
hostinfo.logger(l).Error("Dropping outbound packet, out buffer not large enough", "outCap", cap(out), "payloadLen", len(p), "relayed", useRelay)
}
return
}
fullOut := out
if useRelay {
// Save a header's worth of data at the front of the 'out' buffer.
out = out[header.Len:]
out = out[header.Len:header.Len]
}
if noiseutil.EncryptLockNeeded {
@@ -582,19 +595,23 @@ func (f *Interface) sendNoMetrics(t header.MessageType, st header.MessageSubType
if remote.IsValid() {
err = f.writers[q].WriteTo(out, remote)
if err != nil {
hostinfo.logger(f.l).Error("Failed to write outgoing packet",
if l, ok := logging.RateLimited(f.l, f.udpTxErrLimiter); ok {
hostinfo.logger(l).Error("Failed to write outgoing packet",
"error", err,
"udpAddr", remote,
)
}
}
} else if hr := hostinfo.GetRemote(); hr.IsValid() {
err = f.writers[q].WriteTo(out, hr)
if err != nil {
hostinfo.logger(f.l).Error("Failed to write outgoing packet",
if l, ok := logging.RateLimited(f.l, f.udpTxErrLimiter); ok {
hostinfo.logger(l).Error("Failed to write outgoing packet",
"error", err,
"udpAddr", hr,
)
}
}
} else {
// Try to send via a relay
for _, relayIP := range hostinfo.relayState.CopyRelayIps() {
+31 -23
View File
@@ -24,6 +24,7 @@ import (
"github.com/slackhq/nebula/overlay"
"github.com/slackhq/nebula/overlay/batch"
"github.com/slackhq/nebula/overlay/tio"
"github.com/slackhq/nebula/ratelimit"
"github.com/slackhq/nebula/udp"
)
@@ -124,6 +125,10 @@ type Interface struct {
batchers []*batch.MultiCoalescer
wg sync.WaitGroup
udpTxErrLimiter *ratelimit.Limiter // udp direct transmit error log limiter
udpRelayTxErrLimiter *ratelimit.Limiter // udp relay transmit error log limiter
txOversizeLimiter *ratelimit.Limiter // oversized outbound packet drop log limiter
// fatalErr holds the first unexpected reader error that caused shutdown.
// nil means "no fatal error" (yet)
fatalErr atomic.Pointer[error]
@@ -198,6 +203,19 @@ func NewInterface(ctx context.Context, c *InterfaceConfig) (*Interface, error) {
return nil, errors.New("no connection manager")
}
udpTxErrLimiter, err := ratelimit.New(1, time.Second)
if err != nil {
return nil, err
}
udpRelayTxErrLimiter, err := ratelimit.New(1, time.Second)
if err != nil {
return nil, err
}
txOversizeLimiter, err := ratelimit.New(1, time.Second)
if err != nil {
return nil, err
}
if c.routines <= 1 {
c.PinThreads = false //pinning is not useful unless there's more than one tun reader
}
@@ -230,6 +248,9 @@ func NewInterface(ctx context.Context, c *InterfaceConfig) (*Interface, error) {
conntrackCacheTimeout: c.ConntrackCacheTimeout,
cpuAffinity: c.CpuAffinity,
pinThreads: c.PinThreads,
udpTxErrLimiter: udpTxErrLimiter,
udpRelayTxErrLimiter: udpRelayTxErrLimiter,
txOversizeLimiter: txOversizeLimiter,
metricHandshakes: metrics.GetOrRegisterHistogram("handshakes", nil, metrics.NewExpDecaySample(1028, 0.015)),
metricTxDropped: metrics.GetOrRegisterCounter("udp.tx.dropped", nil),
@@ -488,10 +509,11 @@ func (f *Interface) flushSendBatch(sb *batch.SendBatch, q int) {
func (f *Interface) RegisterConfigChangeCallbacks(c *config.C) {
c.RegisterReloadCallback(f.reloadFirewall)
c.RegisterReloadCallback(f.reloadSendRecvError)
c.RegisterReloadCallback(f.reloadAcceptRecvError)
c.RegisterReloadCallback(f.reloadDisconnectInvalid)
c.RegisterReloadCallback(f.reloadMisc)
for _, reload := range []func(*config.C){
f.reloadSendRecvError, f.reloadAcceptRecvError, f.reloadDisconnectInvalid, f.reloadMisc,
} {
c.RegisterReloadCallback(func(c *config.C, _ *config.Reload) { reload(c) })
}
for _, udpConn := range f.writers {
c.RegisterReloadCallback(udpConn.ReloadConfig)
@@ -508,7 +530,7 @@ func (f *Interface) reloadDisconnectInvalid(c *config.C) {
}
}
func (f *Interface) reloadFirewall(c *config.C) {
func (f *Interface) reloadFirewall(c *config.C, r *config.Reload) {
cs := f.pki.getCertState()
curCert := cs.getCertificate(cert.Version2)
if curCert == nil {
@@ -527,6 +549,7 @@ func (f *Interface) reloadFirewall(c *config.C) {
fw, err := NewFirewallFromConfig(f.l, cs, c)
if err != nil {
f.l.Error("Error while creating firewall during reload", "error", err)
r.Failed()
return
}
@@ -624,17 +647,15 @@ func (f *Interface) reloadMisc(c *config.C) {
}
}
func (f *Interface) emitStats(ctx context.Context, i time.Duration) {
ticker := time.NewTicker(i)
defer ticker.Stop()
// statsEmitter updates the interface's gauges, the stats server runs it on its capture loop.
func (f *Interface) statsEmitter() func() {
udpStats := udp.NewUDPStatsEmitter(f.writers)
certExpirationGauge := metrics.GetOrRegisterGauge("certificate.ttl_seconds", nil)
certInitiatingVersion := metrics.GetOrRegisterGauge("certificate.initiating_version", nil)
certMaxVersion := metrics.GetOrRegisterGauge("certificate.max_version", nil)
emit := func() {
return func() {
f.firewall.EmitStats()
f.handshakeManager.EmitStats()
udpStats()
@@ -651,19 +672,6 @@ func (f *Interface) emitStats(ctx context.Context, i time.Duration) {
certMaxVersion.Update(int64(certState.v1Cert.Version()))
}
}
// Prime gauges so a Prometheus scrape that lands before the first tick
// sees real values instead of the zero defaults (issue #907).
emit()
for {
select {
case <-ctx.Done():
return
case <-ticker.C:
emit()
}
}
}
func (f *Interface) GetHostInfo(vpnIp netip.Addr) *HostInfo {
+5 -14
View File
@@ -3,7 +3,6 @@
package nebula
import (
"context"
"net/netip"
"testing"
"time"
@@ -18,11 +17,8 @@ import (
"github.com/stretchr/testify/require"
)
// Test_emitStats_primesGauges covers issue #907: a Prometheus scrape that
// landed before the first ticker fire used to read 0 for the cert gauges.
// emitStats now primes the gauges before entering the ticker loop. We assert
// the gauge is zero before the first call and non-zero after.
func Test_emitStats_primesGauges(t *testing.T) {
// Priming before the first tick (issue #907) is Start's job, see TestStatsServer_Start_primes.
func Test_statsEmitter_setsGauges(t *testing.T) {
defer metrics.DefaultRegistry.UnregisterAll()
l := test.NewLogger()
@@ -56,17 +52,12 @@ func Test_emitStats_primesGauges(t *testing.T) {
ifce.pki.cs.Store(cs)
ttlGauge := metrics.GetOrRegisterGauge("certificate.ttl_seconds", nil)
require.Zero(t, ttlGauge.Value(), "gauge should be zero before emitStats runs")
require.Zero(t, ttlGauge.Value(), "gauge should be zero before the emitter runs")
// Pre-cancel the context so emitStats returns after priming the gauges
// without ever reading from ticker.C. The one hour interval is just a
// belt-and-suspenders, the test does not expect the ticker to fire.
ctx, cancel := context.WithCancel(context.Background())
cancel()
ifce.emitStats(ctx, time.Hour)
ifce.statsEmitter()()
ttl := ttlGauge.Value()
assert.Positive(t, ttl, "ttl gauge should be primed by emitStats before its first tick")
assert.Positive(t, ttl, "ttl gauge should be set by one run of the emitter")
assert.LessOrEqual(t, ttl, int64(3600))
assert.Equal(t, int64(cert.Version1), metrics.GetOrRegisterGauge("certificate.initiating_version", nil).Value())
assert.Equal(t, int64(cert.Version1), metrics.GetOrRegisterGauge("certificate.max_version", nil).Value())
+26 -4
View File
@@ -65,10 +65,10 @@ func TestReloadFirewall_CertUnsafeNetworksChanged(t *testing.T) {
pki.cs.Store(&CertState{v2Cert: c2, initiatingVersion: cert.Version2})
// Reload with the same YAML so HasChanged("firewall") reports false.
require.NoError(t, cfg.ReloadConfigString(rawYAML))
require.NoError(t, cfg.ReloadConfigString(rawYAML).Err)
require.False(t, cfg.HasChanged("firewall"))
f.reloadFirewall(cfg)
f.reloadFirewall(cfg, nil)
assert.NotSame(t, fw, f.firewall, "firewall pointer should have been replaced")
assert.Equal(t, newUnsafe, f.firewall.unsafeNetworks)
@@ -113,8 +113,30 @@ func TestReloadFirewall_NoChange(t *testing.T) {
l: l,
}
require.NoError(t, cfg.ReloadConfigString(rawYAML))
f.reloadFirewall(cfg)
require.NoError(t, cfg.ReloadConfigString(rawYAML).Err)
f.reloadFirewall(cfg, nil)
assert.Same(t, fw, f.firewall, "firewall should not have been replaced")
}
// Rules that don't parse keep the old firewall, and the reload says it could not apply them
func TestReloadFirewall_BadRulesFailTheReload(t *testing.T) {
l := test.NewLogger()
pki := &PKI{}
pki.cs.Store(&CertState{v2Cert: &dummyCert{version: cert.Version2, networks: []netip.Prefix{netip.MustParsePrefix("10.0.0.1/24")}}, initiatingVersion: cert.Version2})
rules := func(proto string) string {
return "firewall:\n outbound:\n - port: any\n proto: any\n host: any\n inbound:\n - port: any\n proto: " + proto + "\n host: any\n"
}
cfg := config.NewC(l)
require.NoError(t, cfg.LoadString(rules("any")))
fw, err := NewFirewallFromConfig(l, pki.getCertState(), cfg)
require.NoError(t, err)
f := &Interface{pki: pki, firewall: fw, l: l}
cfg.RegisterReloadCallback(f.reloadFirewall)
res := cfg.ReloadConfigString(rules("nope"))
require.NoError(t, res.Err)
assert.True(t, res.Failed)
assert.Same(t, fw, f.firewall)
}
+38 -9
View File
@@ -15,6 +15,7 @@ import (
"time"
"github.com/gaissmai/bart"
"github.com/rcrowley/go-metrics"
"github.com/slackhq/nebula/cert"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/header"
@@ -81,6 +82,8 @@ type LightHouse struct {
updateTrigger chan struct{}
queryChan chan netip.Addr
// queryDropped counts queries QueryServer dropped because queryChan was full.
queryDropped metrics.Counter
calculatedRemotes atomic.Pointer[bart.Table[[]*calculatedRemote]] // Maps VpnAddr to []*calculatedRemote
@@ -116,7 +119,8 @@ func NewLightHouseFromConfig(ctx context.Context, l *slog.Logger, c *config.C, c
nebulaPort: nebulaPort,
punchy: p,
updateTrigger: make(chan struct{}, 1),
queryChan: make(chan netip.Addr, c.GetUint32("handshakes.query_buffer", 64)),
queryChan: make(chan netip.Addr, c.GetUint32("handshakes.query_buffer", 1024)),
queryDropped: metrics.GetOrRegisterCounter("lighthouse.query.dropped", nil),
l: l,
}
h.localAddrsFn = func(al *LocalAllowList) []netip.Addr {
@@ -134,18 +138,21 @@ func NewLightHouseFromConfig(ctx context.Context, l *slog.Logger, c *config.C, c
h.metrics = newLighthouseMetrics()
}
err := h.reload(c, true)
err := h.reload(c, true, nil)
if err != nil {
return nil, err
}
c.RegisterReloadCallback(func(c *config.C) {
err := h.reload(c, false)
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
r.RestartRequiredIfChanged("lighthouse.am_lighthouse", "listen.port", "handshakes.query_buffer", "stats.lighthouse_metrics")
err := h.reload(c, false, r)
switch v := err.(type) {
case *util.ContextualError:
v.Log(l)
r.Failed()
case error:
l.Error("failed to reload lighthouse", "error", err)
r.Failed()
}
})
@@ -186,7 +193,8 @@ func (lh *LightHouse) GetUpdateInterval() int64 {
return lh.interval.Load()
}
func (lh *LightHouse) reload(c *config.C, initial bool) error {
// r is nil on the initial load
func (lh *LightHouse) reload(c *config.C, initial bool, r *config.Reload) error {
if initial || c.HasChanged("lighthouse.advertise_addrs") {
rawAdvAddrs := c.GetStringSlice("lighthouse.advertise_addrs", []string{})
advAddrs := make([]netip.AddrPort, 0)
@@ -221,6 +229,7 @@ func (lh *LightHouse) reload(c *config.C, initial bool) error {
"addr", rawAddr,
"entry", i+1,
)
r.Failed()
continue
}
@@ -355,7 +364,13 @@ func (lh *LightHouse) reload(c *config.C, initial bool) error {
}
}
if initial || c.HasChanged("relay.relays") {
// am_relay decides whether relay.relays applies to us at all
if initial || c.HasChanged("relay.relays") || c.HasChanged("relay.am_relay") {
var before []netip.Addr
if !initial {
before = lh.GetRelaysForMe()
}
switch c.GetBool("relay.am_relay", false) {
case true:
// Relays aren't allowed to specify other relays
@@ -373,6 +388,7 @@ func (lh *LightHouse) reload(c *config.C, initial bool) error {
"relay", v,
"error", err,
)
r.Failed()
} else {
lh.l.Info("Read relay from config", "relay", v)
relaysForMe = append(relaysForMe, configRIP)
@@ -380,6 +396,11 @@ func (lh *LightHouse) reload(c *config.C, initial bool) error {
}
lh.relaysForMe.Store(&relaysForMe)
}
// Tell the lighthouses now rather than at the next interval, but only when what we advertise changed
if !initial && !slices.Equal(before, lh.GetRelaysForMe()) {
lh.TriggerUpdate()
}
}
return nil
@@ -512,14 +533,19 @@ func (lh *LightHouse) Query(vpnAddr netip.Addr) *RemoteList {
return nil
}
// QueryServer is asynchronous so no reply should be expected
// QueryServer is asynchronous and non-blocking, no reply should be expected.
// If the query channel is full, the query is dropped.
func (lh *LightHouse) QueryServer(vpnAddr netip.Addr) {
// Don't put lighthouse addrs in the query channel because we can't query lighthouses about lighthouses
if lh.amLighthouse || lh.IsLighthouseAddr(vpnAddr) {
return
}
lh.queryChan <- vpnAddr
select {
case lh.queryChan <- vpnAddr:
default:
lh.queryDropped.Inc(1)
}
}
func (lh *LightHouse) QueryCache(vpnAddrs []netip.Addr) *RemoteList {
@@ -1235,10 +1261,13 @@ func (lhh *LightHouseHandler) handleHostQuery(n *NebulaMeta, fromVpnAddrs []neti
// sendHostPunchNotification signals the other side to punch some zero byte udp packets
func (lhh *LightHouseHandler) sendHostPunchNotification(n *NebulaMeta, fromVpnAddrs []netip.Addr, punchNotifDest netip.Addr, w EncWriter) {
whereToPunch := fromVpnAddrs[0]
// Resolve the target before taking the querier's RemoteList lock below. GetHostInfo takes the
// hostmap lock, and the hostmap is taken before the lighthouse, which is taken before a
// RemoteList, so looking it up inside the callback would deadlock.
targetHI := lhh.lh.ifce.GetHostInfo(punchNotifDest)
found, ln, err := lhh.lh.queryAndPrepMessage(whereToPunch, func(c *cache) (int, error) {
n = lhh.resetMeta()
n.Type = NebulaMeta_HostPunchNotification
targetHI := lhh.lh.ifce.GetHostInfo(punchNotifDest)
var useVersion cert.Version
if targetHI == nil {
useVersion = lhh.lh.ifce.GetCertState().initiatingVersion
+257 -10
View File
@@ -6,13 +6,16 @@ import (
"errors"
"fmt"
"net/netip"
"sync"
"testing"
"time"
"github.com/gaissmai/bart"
"github.com/slackhq/nebula/cert"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/header"
"github.com/slackhq/nebula/test"
"github.com/slackhq/nebula/udp"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
"go.yaml.in/yaml/v3"
@@ -84,15 +87,15 @@ func TestReloadLighthouseInterval(t *testing.T) {
lh.ifce = &mockEncWriter{}
// The first one routine is kicked off by main.go currently, lets make sure that one dies
require.NoError(t, c.ReloadConfigString("lighthouse:\n interval: 5"))
require.NoError(t, c.ReloadConfigString("lighthouse:\n interval: 5").Err)
assert.Equal(t, int64(5), lh.interval.Load())
// Subsequent calls are killed off by the LightHouse.Reload function
require.NoError(t, c.ReloadConfigString("lighthouse:\n interval: 10"))
require.NoError(t, c.ReloadConfigString("lighthouse:\n interval: 10").Err)
assert.Equal(t, int64(10), lh.interval.Load())
// If this completes then nothing is stealing our reload routine
require.NoError(t, c.ReloadConfigString("lighthouse:\n interval: 11"))
require.NoError(t, c.ReloadConfigString("lighthouse:\n interval: 11").Err)
assert.Equal(t, int64(11), lh.interval.Load())
}
@@ -293,7 +296,7 @@ func TestLighthouse_reload(t *testing.T) {
require.NoError(t, err)
c.ReloadConfigString(string(rc))
err = lh.reload(c, false)
err = lh.reload(c, false, nil)
require.NoError(t, err)
}
@@ -331,7 +334,7 @@ func TestLighthouse_reloadStaticHostMap(t *testing.T) {
}
rc, err := yaml.Marshal(nc)
require.NoError(t, err)
require.NoError(t, c.ReloadConfigString(string(rc)))
require.NoError(t, c.ReloadConfigString(string(rc)).Err)
rl := lh.Query(staticHost)
require.NotNil(t, rl)
@@ -347,7 +350,7 @@ func TestLighthouse_reloadStaticHostMap(t *testing.T) {
}
rc, err = yaml.Marshal(nc)
require.NoError(t, err)
require.NoError(t, c.ReloadConfigString(string(rc)))
require.NoError(t, c.ReloadConfigString(string(rc)).Err)
rl = lh.Query(staticHost)
require.NotNil(t, rl)
@@ -355,7 +358,7 @@ func TestLighthouse_reloadStaticHostMap(t *testing.T) {
assert.Equal(t, []netip.AddrPort{netip.MustParseAddrPort("1.1.1.1:4242")}, rl.CopyAddrs([]netip.Prefix{}))
// Reload with the same config. An unchanged entry must not duplicate.
require.NoError(t, c.ReloadConfigString(string(rc)))
require.NoError(t, c.ReloadConfigString(string(rc)).Err)
rl = lh.Query(staticHost)
require.NotNil(t, rl)
@@ -370,7 +373,7 @@ func TestLighthouse_reloadStaticHostMap(t *testing.T) {
}
rc, err = yaml.Marshal(nc)
require.NoError(t, err)
require.NoError(t, c.ReloadConfigString(string(rc)))
require.NoError(t, c.ReloadConfigString(string(rc)).Err)
// Add a second host alongside the first. Both should be present, neither duplicated.
nc = map[string]any{
@@ -381,7 +384,7 @@ func TestLighthouse_reloadStaticHostMap(t *testing.T) {
}
rc, err = yaml.Marshal(nc)
require.NoError(t, err)
require.NoError(t, c.ReloadConfigString(string(rc)))
require.NoError(t, c.ReloadConfigString(string(rc)).Err)
rl = lh.Query(staticHost)
require.NotNil(t, rl)
@@ -403,7 +406,7 @@ func TestLighthouse_reloadStaticHostMap(t *testing.T) {
}
rc, err = yaml.Marshal(nc)
require.NoError(t, err)
require.NoError(t, c.ReloadConfigString(string(rc)))
require.NoError(t, c.ReloadConfigString(string(rc)).Err)
_, isStatic := lh.GetStaticHostList()[staticHost]
assert.False(t, isStatic)
@@ -795,3 +798,247 @@ func TestLightHouse_logLocalAddrsErr(t *testing.T) {
lh.logLocalAddrsErr(errors.New("something else"))
assert.Contains(t, out.String(), "level=WARN")
}
// heldEncWriter holds the query worker in its first GetHostInfo until release closes.
type heldEncWriter struct {
mockEncWriter
once sync.Once
held chan struct{}
release chan struct{}
}
func (w *heldEncWriter) GetHostInfo(netip.Addr) *HostInfo {
w.once.Do(func() { close(w.held) })
<-w.release
return nil
}
// QueryServer never waits on the query worker, since the packet readers call it and nothing drains the channel at
// shutdown; a query that finds the channel full is dropped and counted. The worker is held on its first query so the
// channel fills.
func TestLighthouse_QueryServerDoesNotBlock(t *testing.T) {
l := test.NewLogger()
cs := testCertState(netip.MustParsePrefix("10.128.0.1/16"))
lh1 := "10.128.0.2"
c := config.NewC(l)
c.Settings["lighthouse"] = map[string]any{"hosts": []any{lh1}}
c.Settings["static_host_map"] = map[string]any{lh1: []any{"1.1.1.1:4242"}}
c.Settings["handshakes"] = map[string]any{"query_buffer": 1}
lh, err := NewLightHouseFromConfig(t.Context(), l, c, cs, nil, nil)
require.NoError(t, err)
w := &heldEncWriter{held: make(chan struct{}), release: make(chan struct{})}
lh.ifce = w
t.Cleanup(func() { close(w.release) })
lh.QueryServer(netip.MustParseAddr("10.128.0.10"))
select {
case <-w.held:
case <-time.After(5 * time.Second):
t.Fatal("the query worker never took the first query")
}
dropped := lh.queryDropped.Count()
done := make(chan struct{})
go func() {
lh.QueryServer(netip.MustParseAddr("10.128.0.11"))
lh.QueryServer(netip.MustParseAddr("10.128.0.12"))
close(done)
}()
select {
case <-done:
case <-time.After(5 * time.Second):
t.Fatal("QueryServer blocked on a full query channel")
}
assert.EqualValues(t, 1, lh.queryDropped.Count()-dropped)
}
// hostmapEncWriter answers GetHostInfo from a real hostmap, the way Interface does.
type hostmapEncWriter struct {
mockEncWriter
hm *HostMap
}
func (w *hostmapEncWriter) GetHostInfo(a netip.Addr) *HostInfo { return w.hm.QueryVpnAddr(a) }
func (w *hostmapEncWriter) GetCertState() *CertState {
return &CertState{initiatingVersion: cert.Version1}
}
// A lighthouse runs four paths that each hold one lock while waiting for the next: CheckAndComplete
// holds the hostmap and takes the handshake manager, StartHandshake holds the handshake manager and
// takes the lighthouse, a host update holds the lighthouse and takes the sender's RemoteList, and a
// punch notification holds the querier's RemoteList while it resolves the target. If that last step
// reaches the hostmap under the RemoteList lock, the four deadlock. Drive them all at once.
func TestLighthouse_PunchNotificationDoesNotCloseLockCycle(t *testing.T) {
l := test.NewLogger()
myVpnNet := netip.MustParsePrefix("10.128.0.1/24")
c := config.NewC(l)
c.Settings["lighthouse"] = map[string]any{"am_lighthouse": true}
c.Settings["listen"] = map[string]any{"port": 4242}
lh, err := NewLightHouseFromConfig(t.Context(), l, c, testCertState(myVpnNet), nil, nil)
require.NoError(t, err)
mainHM := newHostMap(l)
pr := []netip.Prefix{}
mainHM.preferredRanges.Store(&pr)
lh.ifce = &hostmapEncWriter{hm: mainHM}
hm := NewHandshakeManager(l, mainHM, lh, &udp.NoopConn{}, defaultHandshakeConfig)
f := &Interface{hostMap: mainHM, lightHouse: lh, handshakeManager: hm, l: l}
hm.f = f
peer := netip.MustParseAddr("10.128.0.2")
peerUdp := netip.MustParseAddrPort("10.0.0.2:4242")
target := netip.MustParseAddr("10.128.0.3")
targetUdp := netip.MustParseAddrPort("10.0.0.3:4242")
outbound := netip.MustParseAddr("10.128.0.4")
inbound := netip.MustParseAddr("10.128.0.5")
seed := lh.NewRequestHandler()
newLHHostUpdate(peerUdp, peer, []netip.AddrPort{peerUdp}, seed)
newLHHostUpdate(targetUdp, target, []netip.AddrPort{targetUdp}, seed)
stop := make(chan struct{})
var wg sync.WaitGroup
loop := func(body func()) {
defer wg.Done()
for {
select {
case <-stop:
return
default:
body()
}
}
}
wg.Add(4)
// Outbound handshakes keep starting and being dropped.
go loop(func() {
hm.DeleteHostInfo(hm.StartHandshake(outbound, nil))
})
// Inbound handshakes keep completing and being torn down.
var idx uint32
go loop(func() {
idx++
hi := &HostInfo{
vpnAddrs: []netip.Addr{inbound},
localIndexId: idx,
remoteIndexId: idx,
remotes: lh.QueryCache([]netip.Addr{inbound}),
HandshakePacket: map[uint8][]byte{handshakePacketStage0: {byte(idx)}},
}
if _, err := hm.CheckAndComplete(hi, handshakePacketStage0, f); err == nil {
mainHM.DeleteHostInfo(hi)
}
})
// The peer keeps sending host updates.
updates := lh.NewRequestHandler()
go loop(func() {
newLHHostUpdate(peerUdp, peer, []netip.AddrPort{peerUdp}, updates)
})
// The same peer keeps querying for the target, which ends in a punch notification.
queries := lh.NewRequestHandler()
go loop(func() {
sendLHHostRequest(peerUdp, peer, target, queries, nil)
})
time.Sleep(time.Second)
close(stop)
done := make(chan struct{})
go func() { wg.Wait(); close(done) }()
select {
case <-done:
case <-time.After(5 * time.Second):
t.Fatal("deadlock: the lighthouse paths never finished after stop was signalled")
}
}
func TestLighthouse_reloadRelaysTriggersUpdate(t *testing.T) {
l := test.NewLogger()
c := config.NewC(l)
base := "lighthouse:\n am_lighthouse: true\nlisten:\n port: 4242\n"
require.NoError(t, c.LoadString(base+"relay:\n relays: [10.128.0.5]\n"))
lh, err := NewLightHouseFromConfig(t.Context(), l, c, testCertState(netip.MustParsePrefix("10.128.0.1/24")), nil, nil)
require.NoError(t, err)
assert.Empty(t, lh.updateTrigger)
require.NoError(t, c.ReloadConfigString(base+"relay:\n relays: [10.128.0.6]\n").Err)
assert.Equal(t, []netip.Addr{netip.MustParseAddr("10.128.0.6")}, lh.GetRelaysForMe())
require.Len(t, lh.updateTrigger, 1)
<-lh.updateTrigger
// The config changed but what we advertise did not, a bad entry is skipped
require.NoError(t, c.ReloadConfigString(base+"relay:\n relays: [10.128.0.6, nope]\n").Err)
assert.Equal(t, []netip.Addr{netip.MustParseAddr("10.128.0.6")}, lh.GetRelaysForMe())
assert.Empty(t, lh.updateTrigger)
}
func TestLighthouse_reloadAmRelayAlone(t *testing.T) {
l := test.NewLogger()
c := config.NewC(l)
base := "lighthouse:\n am_lighthouse: true\nlisten:\n port: 4242\nrelay:\n relays: [10.128.0.5]\n"
require.NoError(t, c.LoadString(base+" am_relay: false\n"))
lh, err := NewLightHouseFromConfig(t.Context(), l, c, testCertState(netip.MustParsePrefix("10.128.0.1/24")), nil, nil)
require.NoError(t, err)
assert.Equal(t, []netip.Addr{netip.MustParseAddr("10.128.0.5")}, lh.GetRelaysForMe())
// Relays aren't allowed to have relays, flipping am_relay alone must drop them and tell the lighthouses
require.NoError(t, c.ReloadConfigString(base+" am_relay: true\n").Err)
assert.Empty(t, lh.GetRelaysForMe())
require.Len(t, lh.updateTrigger, 1)
<-lh.updateTrigger
require.NoError(t, c.ReloadConfigString(base+" am_relay: false\n").Err)
assert.Equal(t, []netip.Addr{netip.MustParseAddr("10.128.0.5")}, lh.GetRelaysForMe())
require.Len(t, lh.updateTrigger, 1)
}
func TestLighthouse_reloadRestartRequired(t *testing.T) {
l := test.NewLogger()
c := config.NewC(l)
base := "lighthouse:\n am_lighthouse: true\n"
require.NoError(t, c.LoadString(base+"listen:\n port: 4242\n"))
_, err := NewLightHouseFromConfig(t.Context(), l, c, testCertState(netip.MustParsePrefix("10.128.0.1/24")), nil, nil)
require.NoError(t, err)
res := c.ReloadConfigString(base + "listen:\n port: 4242\n")
require.NoError(t, res.Err)
assert.False(t, res.RestartRequired)
res = c.ReloadConfigString(base + "listen:\n port: 4243\n")
require.NoError(t, res.Err)
assert.True(t, res.RestartRequired)
}
// A relay that doesn't parse is skipped and the rest are used, the reload says one was skipped
func TestLighthouse_reloadBadRelayFails(t *testing.T) {
l := test.NewLogger()
c := config.NewC(l)
require.NoError(t, c.LoadString("listen:\n port: 4242\n"))
lh, err := NewLightHouseFromConfig(t.Context(), l, c, testCertState(netip.MustParsePrefix("10.128.0.1/24")), nil, nil)
require.NoError(t, err)
res := c.ReloadConfigString("listen:\n port: 4242\nrelay:\n relays: [10.128.0.2, nope]\n")
require.NoError(t, res.Err)
assert.True(t, res.Failed)
assert.Equal(t, []netip.Addr{netip.MustParseAddr("10.128.0.2")}, lh.GetRelaysForMe())
}
// An advertise_addrs entry inside our own network is skipped and the rest are used, the reload says one was skipped
func TestLighthouse_reloadAdvertiseAddrInNetworkFails(t *testing.T) {
l := test.NewLogger()
c := config.NewC(l)
require.NoError(t, c.LoadString("listen:\n port: 4242\n"))
lh, err := NewLightHouseFromConfig(t.Context(), l, c, testCertState(netip.MustParsePrefix("10.128.0.1/24")), nil, nil)
require.NoError(t, err)
res := c.ReloadConfigString("listen:\n port: 4242\nlighthouse:\n advertise_addrs: [\"1.2.3.4:4242\", \"10.128.0.5:4242\"]\n")
require.NoError(t, res.Err)
assert.True(t, res.Failed)
assert.Equal(t, []netip.AddrPort{netip.MustParseAddrPort("1.2.3.4:4242")}, lh.GetAdvertiseAddrs())
}
+31
View File
@@ -0,0 +1,31 @@
package logging
import "log/slog"
// RateLimiter is what RateLimited needs from a rate limiter, such as a *ratelimit.Limiter.
// Allow reports whether this call may proceed, and how many calls were denied since the last one that did.
type RateLimiter interface {
Allow() (denied uint64, ok bool)
}
// RateLimited checks lim before anything is built.
// When allowed it returns l, with a "suppressed" count attached if lim denied calls since its last allowed one.
// A nil lim allows everything.
//
// if l, ok := logging.RateLimited(f.l, f.udpTxErrLimiter); ok {
// hostinfo.logger(l).Error("Failed to write outgoing packet", "error", err)
// }
func RateLimited(l *slog.Logger, lim RateLimiter) (*slog.Logger, bool) {
if lim == nil {
return l, true
}
denied, ok := lim.Allow()
if !ok {
return nil, false
}
if denied > 0 {
return l.With("suppressed", denied), true
}
return l, true
}
+106
View File
@@ -0,0 +1,106 @@
package logging
import (
"bytes"
"errors"
"io"
"log/slog"
"net/netip"
"strings"
"testing"
"time"
"github.com/slackhq/nebula/ratelimit"
)
type fakeLimiter struct {
denied uint64
ok bool
}
func (f fakeLimiter) Allow() (uint64, bool) { return f.denied, f.ok }
func TestRateLimited(t *testing.T) {
var buf bytes.Buffer
base := NewLogger(&buf)
if l, ok := RateLimited(base, fakeLimiter{ok: false}); ok || l != nil {
t.Fatalf("denied call: ok=%v logger=%v", ok, l)
}
if l, ok := RateLimited(base, fakeLimiter{ok: true}); !ok || l != base {
t.Fatal("nothing denied should hand back the same logger")
}
if l, ok := RateLimited(base, nil); !ok || l != base {
t.Fatal("a nil limiter should allow with the same logger")
}
l, ok := RateLimited(base, fakeLimiter{denied: 7, ok: true})
if !ok {
t.Fatal("allowed call denied")
}
l.Info("hello", "k", "v")
if !strings.Contains(buf.String(), "suppressed=7") {
t.Fatalf("want suppressed=7, got %q", buf.String())
}
}
// The benchmarks compare a log line through slog directly against the same line behind RateLimited: logged with
// nothing denied, denied, and logged after denials.
var (
benchErr = errors.New("sendto: no buffer space available")
benchAddr = netip.MustParseAddrPort("192.0.2.1:4242")
)
func benchLogger() *slog.Logger {
return NewLogger(io.Discard).With("localIndex", 1234, "remoteIndex", 5678)
}
func BenchmarkRateLimitedBaseline(b *testing.B) {
lg := benchLogger()
b.ReportAllocs()
for i := 0; i < b.N; i++ {
lg.Error("Failed to write outgoing packet", "error", benchErr, "udpAddr", benchAddr)
}
}
func BenchmarkRateLimitedAllowed(b *testing.B) {
lg, lim := benchLogger(), mustLimiter(b, 1, 0)
b.ReportAllocs()
for i := 0; i < b.N; i++ {
if l, ok := RateLimited(lg, lim); ok {
l.Error("Failed to write outgoing packet", "error", benchErr, "udpAddr", benchAddr)
}
}
}
func BenchmarkRateLimitedDenied(b *testing.B) {
lg, lim := benchLogger(), mustLimiter(b, 1, time.Hour)
lim.Allow()
b.ReportAllocs()
for i := 0; i < b.N; i++ {
if l, ok := RateLimited(lg, lim); ok {
l.Error("Failed to write outgoing packet", "error", benchErr, "udpAddr", benchAddr)
}
}
}
func BenchmarkRateLimitedAfterDenied(b *testing.B) {
lg, lim := benchLogger(), fakeLimiter{denied: 5, ok: true}
b.ReportAllocs()
for i := 0; i < b.N; i++ {
if l, ok := RateLimited(lg, lim); ok {
l.Error("Failed to write outgoing packet", "error", benchErr, "udpAddr", benchAddr)
}
}
}
func mustLimiter(b *testing.B, burst int, every time.Duration) *ratelimit.Limiter {
l, err := ratelimit.New(burst, every)
if err != nil {
b.Fatal(err)
}
return l
}
+25 -7
View File
@@ -12,6 +12,7 @@ import (
"strings"
"time"
"github.com/slackhq/nebula/commands"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/cpupick"
"github.com/slackhq/nebula/noiseutil"
@@ -71,14 +72,18 @@ func Main(c *config.C, configTest bool, buildVersion string, l *slog.Logger, dev
}
l.Info("Firewall started", "firewallHashes", fw.GetRuleHashes())
ssh, err := sshd.NewSSHServer(ctx, l.With("subsystem", "sshd"))
// The command registry is shared by every surface that exposes these commands, so a
// command cannot reach one of them without reaching the others.
cmdRegistry := commands.New()
ssh, err := sshd.NewSSHServer(ctx, l.With("subsystem", "sshd"), cmdRegistry)
if err != nil {
return nil, util.ContextualizeIfNeeded("Error while creating SSH server", err)
}
wireSSHReload(l, ssh, c)
var sshStart func()
if c.GetBool("sshd.enabled", false) {
sshStart, err = configSSH(l, ssh, c)
sshStart, err = configSSH(l, ssh, c, nil)
if err != nil {
l.Warn("Failed to configure sshd, ssh debugging will not be available", "error", err)
sshStart = nil
@@ -197,7 +202,7 @@ func Main(c *config.C, configTest bool, buildVersion string, l *slog.Logger, dev
if err != nil {
return nil, util.NewContextualError("Failed to open udp listener", m{"queue": i}, err)
}
udpServer.ReloadConfig(c)
udpServer.ReloadConfig(c, nil)
udpConns[i] = udpServer
// If port is dynamic, discover it before the next pass through the for loop
@@ -300,6 +305,7 @@ func Main(c *config.C, configTest bool, buildVersion string, l *slog.Logger, dev
}
var ifce *Interface
var emitStats func()
if !configTest {
ifce, err = NewInterface(ctx, ifConfig)
if err != nil {
@@ -310,6 +316,16 @@ func Main(c *config.C, configTest bool, buildVersion string, l *slog.Logger, dev
lightHouse.ifce = ifce
ifce.RegisterConfigChangeCallbacks(c)
// These are only read at startup
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
r.RestartRequiredIfChanged(
"routines", "tun.routines", "listen.routines", "firewall.conntrack.routine_cache_timeout",
"listen.port", "listen.host", "stats.message_metrics",
"handshakes.try_interval", "handshakes.retries", "handshakes.trigger_buffer",
"tun.drop_local_broadcast", "tun.drop_multicast",
)
r.RestartRequiredIfChanged(osStartupOnlyKeys...)
})
ifce.reloadDisconnectInvalid(c)
ifce.reloadSendRecvError(c)
ifce.reloadAcceptRecvError(c)
@@ -318,9 +334,12 @@ func Main(c *config.C, configTest bool, buildVersion string, l *slog.Logger, dev
go handshakeManager.Run(ctx)
punchy.Start(ctx, ifce, hostMap, lightHouse)
// On the stats server's loop, so the gauges follow the interval it accepted
emitStats = ifce.statsEmitter()
}
stats, err := newStatsServerFromConfig(ctx, l, c, buildVersion, configTest)
stats, err := newStatsServerFromConfig(ctx, l, c, buildVersion, configTest, emitStats)
if err != nil {
return nil, util.ContextualizeIfNeeded("Failed to start stats emitter", err)
}
@@ -329,15 +348,14 @@ func Main(c *config.C, configTest bool, buildVersion string, l *slog.Logger, dev
return nil, nil
}
go ifce.emitStats(ctx, c.GetDuration("stats.interval", time.Second*10))
attachCommands(l, c, ssh, ifce)
attachCommands(l, c, cmdRegistry, ifce)
networkChanges := udp.NewNetworkChangeMonitor(ctx, l, c)
return &Control{
state: StateReady,
f: ifce,
commands: cmdRegistry,
l: l,
ctx: ctx,
cancel: cancel,
+6
View File
@@ -0,0 +1,6 @@
//go:build linux && !android && !e2e_testing
package nebula
// osStartupOnlyKeys are startup only keys that only matter on linux: udp batch size, udp offloads and thread pinning
var osStartupOnlyKeys = []string{"listen.batch", "listen.udp_offloads", "tun.pin_threads", "tun.pin_threads_key", "tun.cpu_affinity"}
+6
View File
@@ -0,0 +1,6 @@
//go:build !linux || android || e2e_testing
package nebula
// osStartupOnlyKeys is empty, none of the OS specific startup keys matter here
var osStartupOnlyKeys []string
+12 -7
View File
@@ -160,9 +160,7 @@ func (f *Interface) readOutsidePackets(via ViaSender, packet []byte, rxc *rxCont
case header.TestReply:
// No-op, useful for the Roaming and connectionManager side-effects above
case header.TestRequest:
const maxCipherOverhead = 16 //todo we use this too often, needs a real importable const
const maxOverhead = header.Len + header.Len + maxCipherOverhead + maxCipherOverhead
if maxOverhead+len(out) > len(rxc.scratch) {
if header.MaxOverhead+len(out) > len(rxc.scratch) {
// A reply that cannot fit in scratch is dropped no matter the log level.
if f.l.Enabled(context.Background(), slog.LevelDebug) {
hostinfo.logger(f.l).Debug("dropping oversized test request", "payloadLen", len(out), "from", via)
@@ -263,12 +261,17 @@ func (f *Interface) handleOutsideRelayPacket(hostinfo *HostInfo, via ViaSender,
// closeTunnel closes a tunnel locally, it does not send a closeTunnel packet to the remote
func (f *Interface) closeTunnel(hostInfo *HostInfo) {
final := f.hostMap.DeleteHostInfo(hostInfo)
if final {
// Hold the hostmap lock across both the delete and the lighthouse eviction,
// to avoid a race with handshake completion.
f.hostMap.Lock()
defer f.hostMap.Unlock()
if !f.hostMap.unlockedDeleteHostInfo(hostInfo) {
return
}
// We no longer have any tunnels with this vpn addr, clear learned lighthouse state to lower memory usage
f.lightHouse.DeleteVpnAddrs(hostInfo.vpnAddrs)
}
}
// sendCloseTunnel is a helper function to send a proper close tunnel packet to a remote
func (f *Interface) sendCloseTunnel(h *HostInfo) {
@@ -473,7 +476,9 @@ func parseV4(data []byte, incoming bool, fp *firewall.ParsedPacket) error {
func (f *Interface) handleOutsideMessagePacket(hostinfo *HostInfo, messageCounter uint64, out []byte, rxc *rxContext) {
err := newPacket(out, true, rxc.fwPacket)
if err != nil {
hostinfo.logger(f.l).Warn("Error while validating inbound packet", "error", err, "packet", out)
if f.l.Enabled(context.Background(), slog.LevelDebug) {
hostinfo.logger(f.l).Debug("Error while validating inbound packet", "error", err, "packet", out)
}
return
}
+57
View File
@@ -0,0 +1,57 @@
//go:build !e2e_testing
package overlay
import (
"fmt"
"log/slog"
"golang.org/x/sys/windows"
"golang.org/x/sys/windows/registry"
)
var tcpipInterfaceKeys = []string{
`SYSTEM\CurrentControlSet\Services\Tcpip\Parameters\Interfaces`,
`SYSTEM\CurrentControlSet\Services\Tcpip6\Parameters\Interfaces`,
}
// disableDNSRegistration prevents publishing overlay addresses in the host's
// DNS records, where clients outside the overlay cannot reach them.
// Failures are non-fatal so registry access cannot prevent tunnel creation.
func disableDNSRegistration(l *slog.Logger, guid windows.GUID) {
id := guid.String()
for _, base := range tcpipInterfaceKeys {
path := base + `\` + id
if err := setDNSRegistrationValues(path); err != nil {
l.Warn("Failed to disable DNS registration on the tun adapter", "error", err, "key", path)
continue
}
l.Debug("Disabled DNS registration on the tun adapter", "key", path)
}
}
func setDNSRegistrationValues(path string) error {
// The stack may not have created the adapter's interface key yet.
k, _, err := registry.CreateKey(registry.LOCAL_MACHINE, path, registry.SET_VALUE)
if err != nil {
return fmt.Errorf("open HKLM\\%s: %w", path, err)
}
defer k.Close()
// Neither RegistrationEnabled nor DisableDynamicUpdate alone reliably
// prevents registration, so all three settings are required.
values := []struct {
name string
data uint32
}{
{"RegistrationEnabled", 0},
{"DisableDynamicUpdate", 1},
{"MaxNumberOfAddressesToRegister", 0},
}
for _, v := range values {
if err := k.SetDWordValue(v.name, v.data); err != nil {
return fmt.Errorf("set %s: %w", v.name, err)
}
}
return nil
}
+28 -4
View File
@@ -50,7 +50,13 @@ func (r Route) String() string {
func makeRouteTree(l *slog.Logger, routes []Route, allowMTU bool) (*bart.Table[routing.Gateways], error) {
routeTree := new(bart.Table[routing.Gateways])
seen := make(map[netip.Prefix]bool, len(routes))
for _, r := range routes {
if seen[r.Cidr] {
l.Warn("route is listed more than once, only the last entry's via is used", "route", r)
}
seen[r.Cidr] = true
if !allowMTU && r.MTU > 0 {
l.Warn("route MTU is not supported on this platform",
"goos", runtime.GOOS,
@@ -98,7 +104,11 @@ func parseRoutes(c *config.C, networks []netip.Prefix) ([]Route, error) {
mtu, ok := rMtu.(int)
if !ok {
mtu, err = strconv.Atoi(rMtu.(string))
s, ok := rMtu.(string)
if !ok {
return nil, fmt.Errorf("entry %v.mtu in tun.routes is not an integer: %v", i+1, rMtu)
}
mtu, err = strconv.Atoi(s)
if err != nil {
return nil, fmt.Errorf("entry %v.mtu in tun.routes is not an integer: %v", i+1, err)
}
@@ -174,7 +184,11 @@ func parseUnsafeRoutes(c *config.C, networks []netip.Prefix) ([]Route, error) {
if rMtu, ok := m["mtu"]; ok {
mtu, ok = rMtu.(int)
if !ok {
mtu, err = strconv.Atoi(rMtu.(string))
s, ok := rMtu.(string)
if !ok {
return nil, fmt.Errorf("entry %v.mtu in tun.unsafe_routes is not an integer: %v", i+1, rMtu)
}
mtu, err = strconv.Atoi(s)
if err != nil {
return nil, fmt.Errorf("entry %v.mtu in tun.unsafe_routes is not an integer: %v", i+1, err)
}
@@ -192,10 +206,15 @@ func parseUnsafeRoutes(c *config.C, networks []netip.Prefix) ([]Route, error) {
metric, ok := rMetric.(int)
if !ok {
_, err = strconv.ParseInt(rMetric.(string), 10, 32)
s, ok := rMetric.(string)
if !ok {
return nil, fmt.Errorf("entry %v.metric in tun.unsafe_routes is not an integer: %v", i+1, rMetric)
}
n, err := strconv.ParseInt(s, 10, 32)
if err != nil {
return nil, fmt.Errorf("entry %v.metric in tun.unsafe_routes is not an integer: %v", i+1, err)
}
metric = int(n)
}
if metric < 0 || metric > math.MaxInt32 {
@@ -248,10 +267,15 @@ func parseUnsafeRoutes(c *config.C, networks []netip.Prefix) ([]Route, error) {
gatewayWeight, ok := rGatewayWeight.(int)
if !ok {
_, err = strconv.ParseInt(rGatewayWeight.(string), 10, 32)
s, ok := rGatewayWeight.(string)
if !ok {
return nil, fmt.Errorf("entry .weight in tun.unsafe_routes[%v].via[%v] is not an integer", i+1, ig+1)
}
w, err := strconv.ParseInt(s, 10, 32)
if err != nil {
return nil, fmt.Errorf("entry .weight in tun.unsafe_routes[%v].via[%v] is not an integer", i+1, ig+1)
}
gatewayWeight = int(w)
}
if gatewayWeight < 1 || gatewayWeight > math.MaxInt32 {
+116
View File
@@ -1,13 +1,17 @@
package overlay
import (
"bytes"
"fmt"
"net/netip"
"strings"
"testing"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/header"
"github.com/slackhq/nebula/routing"
"github.com/slackhq/nebula/test"
"github.com/slackhq/nebula/udp"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
@@ -402,3 +406,115 @@ func Test_makeMultipathUnsafeRouteTree(t *testing.T) {
routing.CalculateBucketsForGateways(expectedGateways)
assert.ElementsMatch(t, expectedGateways, r)
}
func Test_getMTU(t *testing.T) {
l := test.NewLogger()
c := config.NewC(l)
// unset
assert.Equal(t, DefaultMTU, getMTU(l, c))
// the largest a relayed packet can carry in one underlay buffer
require.Equal(t, udp.MTU, MaxMTU+header.MaxOverhead)
c.Settings["tun"] = map[string]any{"mtu": MaxMTU}
assert.Equal(t, MaxMTU, getMTU(l, c))
// too big is capped
c.Settings["tun"] = map[string]any{"mtu": 9000}
assert.Equal(t, MaxMTU, getMTU(l, c))
}
func Test_getAllRoutesFromConfigCapsMTU(t *testing.T) {
l := test.NewLogger()
c := config.NewC(l)
c.Settings["tun"] = map[string]any{
"routes": []any{
map[string]any{"mtu": "9000", "route": "10.0.0.0/29"},
map[string]any{"mtu": "1300", "route": "10.0.0.8/29"},
},
"unsafe_routes": []any{
map[string]any{"via": "127.0.0.1", "mtu": "9000", "route": "1.0.0.0/29"},
},
}
changed, routes, err := getAllRoutesFromConfig(l, c, []netip.Prefix{netip.MustParsePrefix("10.0.0.0/24")}, true)
require.NoError(t, err)
assert.True(t, changed)
mtus := map[string]int{}
for _, r := range routes {
mtus[r.Cidr.String()] = r.MTU
}
assert.Equal(t, map[string]int{"10.0.0.0/29": MaxMTU, "10.0.0.8/29": 1300, "1.0.0.0/29": MaxMTU}, mtus)
}
// The route tree keeps one entry per CIDR, so a CIDR listed twice only routes by the last one. Say so instead of
// letting the earlier via, or on Windows the earlier metric, quietly lose
func Test_makeRouteTree_WarnsOnDuplicate(t *testing.T) {
buf := &bytes.Buffer{}
l := test.NewLoggerWithOutput(buf)
via := func(s string) routing.Gateways {
return routing.Gateways{routing.NewGateway(netip.MustParseAddr(s), 1)}
}
cidr := netip.MustParsePrefix("10.251.1.0/24")
_, err := makeRouteTree(l, []Route{
{Cidr: cidr, Via: via("10.250.0.2"), Install: true, Metric: 100},
{Cidr: netip.MustParsePrefix("10.251.2.0/24"), Via: via("10.250.0.2"), Install: true},
}, false)
require.NoError(t, err)
assert.Empty(t, buf.String())
tree, err := makeRouteTree(l, []Route{
{Cidr: cidr, Via: via("10.250.0.2"), Install: true, Metric: 100},
{Cidr: cidr, Via: via("10.250.0.3"), Install: true, Metric: 200},
}, false)
require.NoError(t, err)
assert.Equal(t, 1, strings.Count(buf.String(), "route is listed more than once, only the last entry's via is used"))
gw, ok := tree.Lookup(netip.MustParseAddr("10.251.1.1"))
require.True(t, ok)
assert.Equal(t, netip.MustParseAddr("10.250.0.3"), gw[0].Addr(), "the last entry wins")
}
// A quoted metric or weight is a number written as a string, it means that number. Anything else is an error, not a
// panic
func Test_parseUnsafeRoutes_QuotedNumbers(t *testing.T) {
c := config.NewC(test.NewLogger())
n := netip.MustParsePrefix("10.0.0.1/24")
route := func(metric, weight any) {
c.Settings["tun"] = map[string]any{"unsafe_routes": []any{map[string]any{
"route": "1.0.0.0/8",
"metric": metric,
"via": []any{map[string]any{"gateway": "10.0.0.2", "weight": weight}},
}}}
}
route("100", "5")
routes, err := parseUnsafeRoutes(c, []netip.Prefix{n})
require.NoError(t, err)
require.Len(t, routes, 1)
assert.Equal(t, 100, routes[0].Metric)
assert.Equal(t, "{addr: 10.0.0.2, weight: 5}", routes[0].Via[0].String())
route(1.5, 1)
_, err = parseUnsafeRoutes(c, []netip.Prefix{n})
require.EqualError(t, err, "entry 1.metric in tun.unsafe_routes is not an integer: 1.5")
route(1, true)
_, err = parseUnsafeRoutes(c, []netip.Prefix{n})
require.EqualError(t, err, "entry .weight in tun.unsafe_routes[1].via[1] is not an integer")
}
// An mtu that is neither a number nor a quoted one is an error, not a panic
func Test_parseRoutes_MTUNotANumber(t *testing.T) {
c := config.NewC(test.NewLogger())
n := netip.MustParsePrefix("10.0.0.1/24")
c.Settings["tun"] = map[string]any{"routes": []any{map[string]any{"route": "1.0.0.0/8", "mtu": 1500.5}}}
_, err := parseRoutes(c, []netip.Prefix{n})
require.EqualError(t, err, "entry 1.mtu in tun.routes is not an integer: 1500.5")
c.Settings["tun"] = map[string]any{"unsafe_routes": []any{map[string]any{"route": "1.0.0.0/8", "via": "10.0.0.2", "mtu": 1500.5}}}
_, err = parseUnsafeRoutes(c, []netip.Prefix{n})
require.EqualError(t, err, "entry 1.mtu in tun.unsafe_routes is not an integer: 1500.5")
}
+31 -2
View File
@@ -7,11 +7,27 @@ import (
"net/netip"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/header"
"github.com/slackhq/nebula/udp"
"github.com/slackhq/nebula/util"
)
const DefaultMTU = 1300
// MaxMTU is the largest tun.mtu, or route mtu, the underlay can carry.
// Every underlay buffer is udp.MTU bytes, and a relayed packet adds header.MaxOverhead to its plaintext.
const MaxMTU = udp.MTU - header.MaxOverhead
// getMTU reads tun.mtu, capping a value too big for the underlay to carry.
func getMTU(l *slog.Logger, c *config.C) int {
mtu := c.GetInt("tun.mtu", DefaultMTU)
if mtu > MaxMTU {
l.Warn("tun.mtu is too big for the underlay to carry, capping it", "mtu", mtu, "maxMTU", MaxMTU)
return MaxMTU
}
return mtu
}
type NameError struct {
Name string
Underlying error
@@ -25,9 +41,16 @@ func (e *NameError) Error() string {
type DeviceFactory func(c *config.C, l *slog.Logger, vpnNetworks []netip.Prefix, routines int) (Device, error)
func NewDeviceFromConfig(c *config.C, l *slog.Logger, vpnNetworks []netip.Prefix, routines int) (Device, error) {
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
r.RestartRequiredIfChanged("tun.disabled")
})
switch {
case c.GetBool("tun.disabled", false):
tun := newDisabledTun(vpnNetworks, c.GetInt("tun.tx_queue", 500), c.GetBool("stats.message_metrics", false), l)
tun := newDisabledTun(vpnNetworks, c.GetInt("tun.tx_queue", 5000), c.GetBool("stats.message_metrics", false), l)
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
r.RestartRequiredIfChanged("tun.tx_queue", "stats.message_metrics")
})
return tun, nil
default:
@@ -41,7 +64,7 @@ func NewFdDeviceFromConfig(fd *int) DeviceFactory {
}
}
func getAllRoutesFromConfig(c *config.C, vpnNetworks []netip.Prefix, initial bool) (bool, []Route, error) {
func getAllRoutesFromConfig(l *slog.Logger, c *config.C, vpnNetworks []netip.Prefix, initial bool) (bool, []Route, error) {
if !initial && !c.HasChanged("tun.routes") && !c.HasChanged("tun.unsafe_routes") {
return false, nil, nil
}
@@ -57,6 +80,12 @@ func getAllRoutesFromConfig(c *config.C, vpnNetworks []netip.Prefix, initial boo
}
routes = append(routes, unsafeRoutes...)
for i, r := range routes {
if r.MTU > MaxMTU {
l.Warn("route mtu is too big for the underlay to carry, capping it", "route", r.Cidr, "mtu", r.MTU, "maxMTU", MaxMTU)
routes[i].MTU = MaxMTU
}
}
return true, routes, nil
}
+6 -5
View File
@@ -38,16 +38,17 @@ func newTunFromFd(c *config.C, l *slog.Logger, deviceFd int, vpnNetworks []netip
l: l,
}
err := t.reload(c, true)
err := t.reload(c, true, nil)
if err != nil {
_ = file.Close()
return nil, err
}
c.RegisterReloadCallback(func(c *config.C) {
err := t.reload(c, false)
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
err := t.reload(c, false, r)
if err != nil {
util.LogWithContextIfNeeded("failed to reload tun device", err, t.l)
r.Failed()
}
})
@@ -67,8 +68,8 @@ func (t *tun) Activate() error {
return nil
}
func (t *tun) reload(c *config.C, initial bool) error {
change, routes, err := getAllRoutesFromConfig(c, t.vpnNetworks, initial)
func (t *tun) reload(c *config.C, initial bool, r *config.Reload) error {
change, routes, err := getAllRoutesFromConfig(t.l, c, t.vpnNetworks, initial)
if err != nil {
return err
}
+45
View File
@@ -0,0 +1,45 @@
//go:build (openbsd || netbsd) && !e2e_testing
package overlay
import (
"errors"
"fmt"
"log/slog"
"os"
"regexp"
"runtime"
"golang.org/x/sys/unix"
)
var deviceNameRE = regexp.MustCompile(`^tun[0-9]+$`)
// tunDevNeedsRestart is true when tun.dev asks for a particular tun that isn't the running one. Anything else takes the
// first free one, and the one we have will do
func tunDevNeedsRestart(name, running string) bool {
return deviceNameRE.MatchString(name) && name != running
}
// openTun opens /dev/name, or the first /dev/tunN nothing else has open when name is empty or not tunN
func openTun(l *slog.Logger, name string) (int, string, error) {
if name != "" && !deviceNameRE.MatchString(name) {
// The name doesn't mean anything here, so like darwin a bad one isn't worth refusing to start over
l.Warn("interface name must be tun[0-9]+ on "+runtime.GOOS+", ignoring", "name", name)
name = ""
}
if name != "" {
fd, err := unix.Open("/dev/"+name, os.O_RDWR, 0)
return fd, name, err
}
for i := 0; ; i++ {
name = fmt.Sprintf("tun%d", i)
fd, err := unix.Open("/dev/"+name, os.O_RDWR, 0)
if err == nil {
return fd, name, nil
}
if !errors.Is(err, unix.EBUSY) {
return -1, "", fmt.Errorf("no free tun device, %s: %w", name, err)
}
}
}
+80
View File
@@ -0,0 +1,80 @@
//go:build (openbsd || netbsd) && !e2e_testing
package overlay
import (
"net/netip"
"testing"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/test"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
// A tun.dev that isn't tunN is ignored like darwin does, and the first free tun is used. Only naming a different tun
// asks for a restart
func TestTunReloadDevRestart(t *testing.T) {
newWithDev := func(dev string) (*config.C, *tun, error) {
c := config.NewC(test.NewLogger())
require.NoError(t, c.LoadString("tun:\n dev: "+dev+"\n"))
tn, err := newTun(c, test.NewLogger(), []netip.Prefix{netip.MustParsePrefix("10.1.0.2/24")}, false)
if err == nil {
t.Cleanup(func() { _ = tn.Close() })
}
return c, tn, err
}
if _, _, err := newWithDev(""); err != nil {
t.Skipf("can't create a tun device here: %v", err)
}
c, tn, err := newWithDev("nope")
require.NoError(t, err, "a bad name should fall back to a free tun")
require.Regexp(t, `^tun[0-9]+$`, tn.Device)
for _, dev := range []string{"nope", "", tn.Device, "nope"} {
res := c.ReloadConfigString("tun:\n dev: " + dev + "\n")
require.NoError(t, res.Err)
assert.False(t, res.Failed, dev)
assert.False(t, res.RestartRequired, dev)
}
other := "tun0"
if tn.Device == other {
other = "tun1"
}
res := c.ReloadConfigString("tun:\n dev: " + other + "\n")
require.NoError(t, res.Err)
assert.Equal(t, []string{"tun.dev"}, res.RestartKeys)
}
// An empty tun.dev takes the first tun nothing else has open. Reloading it empty, or naming the one we have, needs
// no restart
func TestTunEmptyDevPicksAFreeOne(t *testing.T) {
newEmpty := func() (*config.C, *tun, error) {
c := config.NewC(test.NewLogger())
require.NoError(t, c.LoadString("tun: {}\n"))
tn, err := newTun(c, test.NewLogger(), []netip.Prefix{netip.MustParsePrefix("10.1.0.2/24")}, false)
if err == nil {
t.Cleanup(func() { _ = tn.Close() })
}
return c, tn, err
}
c, first, err := newEmpty()
if err != nil {
t.Skipf("can't create a tun device here: %v", err)
}
_, second, err := newEmpty()
require.NoError(t, err, "the second tun should skip the one the first has open")
assert.NotEqual(t, first.Device, second.Device)
for _, cfg := range []string{
"tun: {}\npunchy:\n punch: true\n",
"tun:\n dev: " + first.Device + "\n",
"tun: {}\n",
} {
res := c.ReloadConfigString(cfg)
require.NoError(t, res.Err)
assert.False(t, res.Failed)
assert.False(t, res.RestartRequired, cfg)
}
}
+3
View File
@@ -10,6 +10,9 @@ import (
"github.com/slackhq/nebula/wfp"
)
// interfaceBypassKeys are only read when the interface bypass is installed
var interfaceBypassKeys = []string{"tun.windows_bypass_wdf"}
// installInterfaceBypass installs a WFP PERMIT filter scoped to the wintun interface LUID so inbound traffic on the
// nebula adapter bypasses Windows Defender Firewall.
func installInterfaceBypass(l *slog.Logger, luid uint64) closer {
+3
View File
@@ -4,6 +4,9 @@ package overlay
import "log/slog"
// interfaceBypassKeys is empty, windows-386 has no bypass
var interfaceBypassKeys []string
// installInterfaceBypass is a no-op on windows-386 because we don't currently build for it.
func installInterfaceBypass(_ *slog.Logger, _ uint64) closer {
return nil
+149 -38
View File
@@ -6,8 +6,10 @@ import (
"errors"
"fmt"
"log/slog"
"math/bits"
"net/netip"
"os"
"sync"
"sync/atomic"
"syscall"
"unsafe"
@@ -28,7 +30,7 @@ type tun struct {
DefaultMTU int
Routes atomic.Pointer[[]Route]
routeTree atomic.Pointer[bart.Table[routing.Gateways]]
linkAddr *netroute.LinkAddr
linkAddr atomic.Pointer[netroute.LinkAddr]
// hostOwned means the fd arrived from the OS, which has already configured addressing, mtu
// and routes for it. NEPacketTunnelProvider on darwin does this.
hostOwned bool
@@ -78,17 +80,31 @@ type ifreqAlias6 struct {
Lifetime addrLifetime
}
// utunIndex is the utun tun.dev asks for, or -1 for whichever is free. ok is false for a name that isn't utunN, which
// also gets whichever is free
func utunIndex(name string) (index int, ok bool) {
if name == "" || name == "utun" {
return -1, true
}
if _, err := fmt.Sscanf(name, "utun%d", &index); err != nil || index < 0 {
return -1, false
}
return index, true
}
// tunDevNeedsRestart is true when tun.dev asks for a particular utun that isn't the running one. Anything else takes
// whichever is free, and the one we have will do
func tunDevNeedsRestart(name, running string) bool {
index, _ := utunIndex(name)
return index >= 0 && fmt.Sprintf("utun%d", index) != running
}
func newTun(c *config.C, l *slog.Logger, vpnNetworks []netip.Prefix, _ bool) (*tun, error) {
name := c.GetString("tun.dev", "")
ifIndex := -1
if name != "" && name != "utun" {
_, err := fmt.Sscanf(name, "utun%d", &ifIndex)
if err != nil || ifIndex < 0 {
ifIndex, ok := utunIndex(c.GetString("tun.dev", ""))
if !ok {
// NOTE: we don't make this error so we don't break existing
// configs that set a name before it was used.
l.Warn("interface name must be utun[0-9]+ on Darwin, ignoring")
ifIndex = -1
}
}
fd, err := unix.Socket(unix.AF_SYSTEM, unix.SOCK_DGRAM, unix.AF_SYS_CONTROL)
@@ -112,7 +128,7 @@ func newTun(c *config.C, l *slog.Logger, vpnNetworks []netip.Prefix, _ bool) (*t
return nil, fmt.Errorf("SYS_CONNECT: %v", err)
}
name, err = unix.GetsockoptString(fd, unix.AF_SYS_CONTROL, _UTUN_OPT_IFNAME)
name, err := unix.GetsockoptString(fd, unix.AF_SYS_CONTROL, _UTUN_OPT_IFNAME)
if err != nil {
return nil, fmt.Errorf("failed to retrieve tun name: %w", err)
}
@@ -126,19 +142,24 @@ func newTun(c *config.C, l *slog.Logger, vpnNetworks []netip.Prefix, _ bool) (*t
f: os.NewFile(uintptr(fd), ""),
Device: name,
vpnNetworks: vpnNetworks,
DefaultMTU: c.GetInt("tun.mtu", DefaultMTU),
DefaultMTU: getMTU(l, c),
l: l,
}
err = t.reload(c, true)
err = t.reload(c, true, nil)
if err != nil {
return nil, err
}
c.RegisterReloadCallback(func(c *config.C) {
err := t.reload(c, false)
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
r.RestartRequiredIfChanged("tun.mtu")
if tunDevNeedsRestart(c.GetString("tun.dev", ""), t.Device) {
r.RestartRequiredIfChanged("tun.dev")
}
err := t.reload(c, false, r)
if err != nil {
util.LogWithContextIfNeeded("failed to reload tun device", err, t.l)
r.Failed()
}
})
@@ -167,19 +188,20 @@ func newTunFromFd(c *config.C, l *slog.Logger, deviceFd int, vpnNetworks []netip
f: file,
Device: utunNameFromFd(deviceFd),
vpnNetworks: vpnNetworks,
DefaultMTU: c.GetInt("tun.mtu", DefaultMTU),
DefaultMTU: getMTU(l, c),
hostOwned: true,
l: l,
}
if err := t.reload(c, true); err != nil {
if err := t.reload(c, true, nil); err != nil {
_ = file.Close()
return nil, err
}
c.RegisterReloadCallback(func(c *config.C) {
if err := t.reload(c, false); err != nil {
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
if err := t.reload(c, false, r); err != nil {
util.LogWithContextIfNeeded("failed to reload tun device", err, t.l)
r.Failed()
}
})
@@ -243,7 +265,7 @@ func (t *tun) Activate() error {
if linkAddr == nil {
return fmt.Errorf("unable to discover link_addr for tun interface")
}
t.linkAddr = linkAddr
t.linkAddr.Store(linkAddr)
for _, network := range t.vpnNetworks {
if network.Addr().Is4() {
@@ -266,9 +288,10 @@ func (t *tun) Activate() error {
}
// Unsafe path routes
return t.addRoutes(false)
return t.addRoutes(false, nil)
}
// activate4 assigns our vpn address as a /32 and routes the vpn network to the tun explicitly.
func (t *tun) activate4(network netip.Prefix) error {
s, err := unix.Socket(
unix.AF_INET,
@@ -295,22 +318,32 @@ func (t *tun) activate4(network netip.Prefix) error {
MaskAddr: unix.RawSockaddrInet4{
Len: unix.SizeofSockaddrInet4,
Family: unix.AF_INET,
Addr: prefixToMask(network).As4(),
Addr: [4]byte{255, 255, 255, 255}, //always make the address a /32
},
}
// On Darwin, VPNs don't edit the routing table directly.
// A network extension hands its routes to the system, and configd installs them.
// configd replaces any route already holding the same prefix, including ours.
// As an example, Tailscale with an exit node treats any interface prefix wider than one address as a LAN.
// With LAN access off, it routes each LAN into Tailscale's utun.
// With LAN access on, it excludes each LAN from its tunnel.
// configd routes an excluded prefix like its own routes would, ignoring ours, so it goes out the default route.
// Either way our route is replaced, and traffic for the vpn network no longer reaches nebula.
// A single address isn't a LAN to Tailscale, so our route is left alone.
// We route the vpn network to the tun explicitly instead.
if err := ioctl(uintptr(s), unix.SIOCAIFADDR, uintptr(unsafe.Pointer(&ifr))); err != nil {
return fmt.Errorf("failed to set tun v4 address: %s", err)
}
err = addRoute(network, t.linkAddr)
if err != nil {
return err
if err := addRoute(network, t.linkAddr.Load()); err != nil {
return fmt.Errorf("failed to add route for %v: %w", network.Masked(), err)
}
return nil
}
// activate6 is activate4 for IPv6
func (t *tun) activate6(network netip.Prefix) error {
s, err := unix.Socket(
unix.AF_INET6,
@@ -332,25 +365,29 @@ func (t *tun) activate6(network netip.Prefix) error {
PrefixMask: unix.RawSockaddrInet6{
Len: unix.SizeofSockaddrInet6,
Family: unix.AF_INET6,
Addr: prefixToMask(network).As16(),
// Always set the netmask to be a /128. See activate4 for why we do this.
Addr: [16]byte{255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255},
},
Lifetime: addrLifetime{
// never expires
Vltime: 0xffffffff,
Pltime: 0xffffffff,
},
Flags: _IN6_IFF_NODAD,
Flags: _IN6_IFF_NODAD, // No DAD, a utun has no neighbors.
}
if err := ioctl(uintptr(s), _SIOCAIFADDR_IN6, uintptr(unsafe.Pointer(&ifr))); err != nil {
return fmt.Errorf("failed to set tun address: %s", err)
}
if err := addRoute(network, t.linkAddr.Load()); err != nil {
return fmt.Errorf("failed to add route for %v: %w", network.Masked(), err)
}
return nil
}
func (t *tun) reload(c *config.C, initial bool) error {
change, routes, err := getAllRoutesFromConfig(c, t.vpnNetworks, initial)
func (t *tun) reload(c *config.C, initial bool, r *config.Reload) error {
change, routes, err := getAllRoutesFromConfig(t.l, c, t.vpnNetworks, initial)
if err != nil {
return err
}
@@ -370,16 +407,18 @@ func (t *tun) reload(c *config.C, initial bool) error {
if !initial {
// Remove first, if the system removes a wanted route hopefully it will be re-added next
err := t.removeRoutes(findRemovedRoutes(routes, *oldRoutes))
err := t.removeRoutes(findRemovedRoutes(routes, *oldRoutes), r)
if err != nil {
util.LogWithContextIfNeeded("Failed to remove routes", err, t.l)
r.Failed()
}
// Ensure any routes we actually want are installed
err = t.addRoutes(true)
err = t.addRoutes(true, r)
if err != nil {
// Catch any stray logs
util.LogWithContextIfNeeded("Failed to add routes", err, t.l)
r.Failed()
}
}
@@ -394,6 +433,56 @@ func (t *tun) RoutesFor(ip netip.Addr) routing.Gateways {
return routing.Gateways{}
}
func fetchRoutes() ([]netroute.Message, error) {
rib, err := netroute.FetchRIB(unix.AF_UNSPEC, unix.NET_RT_DUMP, 0)
if err != nil {
return nil, err
}
return netroute.ParseRIB(unix.NET_RT_DUMP, rib)
}
// routeInterfaceIn returns the index of the interface prefix is routed through, or 0 if msgs has no route for exactly
// that prefix. Routes scoped to an interface are skipped, they don't stop us adding ours.
func routeInterfaceIn(msgs []netroute.Message, prefix netip.Prefix) int {
for _, msg := range msgs {
m, ok := msg.(*netroute.RouteMessage)
if !ok || len(m.Addrs) <= unix.RTAX_NETMASK || m.Flags&unix.RTF_IFSCOPE != 0 {
continue
}
if p, ok := routePrefix(m.Addrs[unix.RTAX_DST], m.Addrs[unix.RTAX_NETMASK]); ok && p == prefix.Masked() {
return m.Index
}
}
return 0
}
// routePrefix returns the prefix a route covers. Host routes have no netmask.
func routePrefix(dst, mask netroute.Addr) (netip.Prefix, bool) {
switch d := dst.(type) {
case *netroute.Inet4Addr:
bits := 32
if m, ok := mask.(*netroute.Inet4Addr); ok {
bits = maskBits(m.IP[:])
}
return netip.PrefixFrom(netip.AddrFrom4(d.IP), bits), true
case *netroute.Inet6Addr:
bits := 128
if m, ok := mask.(*netroute.Inet6Addr); ok {
bits = maskBits(m.IP[:])
}
return netip.PrefixFrom(netip.AddrFrom16(d.IP), bits), true
}
return netip.Prefix{}, false
}
func maskBits(mask []byte) int {
n := 0
for _, b := range mask {
n += bits.OnesCount8(b)
}
return n
}
// Get the LinkAddr for the interface of the given name
// Is there an easier way to fetch this when we create the interface?
// Maybe SIOCGIFINDEX? but this doesn't appear to exist in the darwin headers.
@@ -422,13 +511,20 @@ func getLinkAddr(name string) (*netroute.LinkAddr, error) {
return nil, nil
}
func (t *tun) addRoutes(logErrors bool) error {
func (t *tun) addRoutes(logErrors bool, rl *config.Reload) error {
// The route tree is still ours, the system routing table is not
if t.hostOwned {
return nil
}
// Activate hasn't run yet, it will install the routes
linkAddr := t.linkAddr.Load()
if linkAddr == nil {
return nil
}
routes := *t.Routes.Load()
systemRoutes := sync.OnceValues(fetchRoutes)
for _, r := range routes {
if len(r.Via) == 0 || !r.Install {
@@ -436,39 +532,54 @@ func (t *tun) addRoutes(logErrors bool) error {
continue
}
err := addRoute(r.Cidr, t.linkAddr)
err := addRoute(r.Cidr, linkAddr)
if err != nil {
if errors.Is(err, unix.EEXIST) {
t.l.Warn("unable to add unsafe_route, identical route already exists", "route", r.Cidr)
// Every route we added already exists on a reload, only warn about someone else's
msgs, err := systemRoutes()
if err != nil {
t.l.Warn("unable to add unsafe_route, a route already exists but its interface could not be determined", "route", r.Cidr, "error", err)
} else if routeInterfaceIn(msgs, r.Cidr) != linkAddr.Index {
t.l.Warn("unable to add unsafe_route, the destination is already routed through another interface", "route", r.Cidr)
}
} else {
retErr := util.NewContextualError("Failed to add route", map[string]any{"route": r}, err)
if logErrors {
retErr.Log(t.l)
rl.Failed()
} else {
return retErr
}
}
} else {
t.l.Info("Added route", "route", r)
// A cidr listed twice would otherwise find our own route missing from an older fetch
systemRoutes = sync.OnceValues(fetchRoutes)
}
}
return nil
}
func (t *tun) removeRoutes(routes []Route) error {
if t.hostOwned {
func (t *tun) removeRoutes(routes []Route, rl *config.Reload) error {
linkAddr := t.linkAddr.Load()
if t.hostOwned || linkAddr == nil {
return nil
}
for _, r := range routes {
if !r.Install {
if len(r.Via) == 0 || !r.Install {
// addRoutes never installs these
continue
}
err := delRoute(r.Cidr, t.linkAddr)
if err != nil {
err := delRoute(r.Cidr, linkAddr)
if errors.Is(err, unix.ESRCH) {
// Already gone, which is what we wanted
t.l.Info("Route was already removed", "route", r)
} else if err != nil {
t.l.Error("Failed to remove route", "error", err, "route", r)
rl.Failed()
} else {
t.l.Info("Removed route", "route", r)
}
+118
View File
@@ -0,0 +1,118 @@
//go:build darwin && !ios && !e2e_testing
package overlay
import (
"net"
"net/netip"
"os"
"os/exec"
"strings"
"testing"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/routing"
"github.com/slackhq/nebula/test"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
netroute "golang.org/x/net/route"
"golang.org/x/sys/unix"
)
// Every mac routes loopback through lo0
func TestRouteInterfaceIn(t *testing.T) {
msgs, err := fetchRoutes()
require.NoError(t, err)
routeInterface := func(s string) int { return routeInterfaceIn(msgs, netip.MustParsePrefix(s)) }
lo0, err := net.InterfaceByName("lo0")
require.NoError(t, err)
assert.Equal(t, lo0.Index, routeInterface("127.0.0.0/8"))
assert.Equal(t, lo0.Index, routeInterface("127.0.0.1/32"), "a host route has no netmask")
assert.Equal(t, lo0.Index, routeInterface("::1/128"))
assert.Equal(t, 0, routeInterface("127.0.0.0/16"), "only an exact prefix counts")
// Most macs also list the default route scoped to their other interfaces
out, err := exec.Command("route", "-n", "get", "default").Output()
require.NoError(t, err)
var name string
for _, line := range strings.Split(string(out), "\n") {
if v, ok := strings.CutPrefix(strings.TrimSpace(line), "interface: "); ok {
name = v
}
}
if name == "" {
t.Skip("no default route")
}
def, err := net.InterfaceByName(name)
require.NoError(t, err)
assert.Equal(t, def.Index, routeInterface("0.0.0.0/0"))
}
func TestRouteInterfaceIn_SkipsScopedCopies(t *testing.T) {
route := func(index, flags int) netroute.Message {
addrs := make([]netroute.Addr, unix.RTAX_MAX)
addrs[unix.RTAX_DST] = &netroute.Inet4Addr{IP: [4]byte{192, 168, 1, 0}}
addrs[unix.RTAX_NETMASK] = &netroute.Inet4Addr{IP: [4]byte{255, 255, 255, 0}}
return &netroute.RouteMessage{Index: index, Flags: flags, Addrs: addrs}
}
p := netip.MustParsePrefix("192.168.1.0/24")
assert.Equal(t, 7, routeInterfaceIn([]netroute.Message{route(4, unix.RTF_UP|unix.RTF_IFSCOPE), route(7, unix.RTF_UP)}, p))
assert.Equal(t, 0, routeInterfaceIn([]netroute.Message{route(4, unix.RTF_UP|unix.RTF_IFSCOPE)}, p))
}
// A reload can happen before Activate
func TestTunRoutesBeforeActivate(t *testing.T) {
tn := &tun{l: test.NewLogger()}
routes := []Route{{
Cidr: netip.MustParsePrefix("10.253.1.0/24"),
Via: routing.Gateways{routing.NewGateway(netip.MustParseAddr("10.250.0.2"), 1)},
Install: true,
}}
tn.Routes.Store(&routes)
assert.NoError(t, tn.addRoutes(true, nil))
assert.NoError(t, tn.removeRoutes(routes, nil))
}
// A tun.routes entry with no via is never installed, so removing it must not delete a route for the same cidr that
// belongs to someone else
func TestTunReloadLeavesRoutesItNeverInstalled(t *testing.T) {
if os.Geteuid() != 0 {
t.Skip("creating a utun and adding routes needs root")
}
lo0, err := net.InterfaceByName("lo0")
require.NoError(t, err)
cidr := "10.250.0.128/25"
require.NoError(t, exec.Command("route", "-q", "-n", "add", "-net", cidr, "-interface", "lo0").Run())
t.Cleanup(func() { _ = exec.Command("route", "-q", "-n", "delete", "-net", cidr).Run() })
l := test.NewLogger()
c := config.NewC(l)
require.NoError(t, c.LoadString("tun:\n routes:\n - route: "+cidr+"\n mtu: 1300\n"))
tn, err := newTun(c, l, []netip.Prefix{netip.MustParsePrefix("10.250.0.1/24")}, false)
require.NoError(t, err)
t.Cleanup(func() { _ = tn.Close() })
require.NoError(t, tn.Activate())
res := c.ReloadConfigString("tun: {}\n")
require.NoError(t, res.Err)
assert.False(t, res.Failed, "removing a route we never installed failed the reload")
msgs, err := fetchRoutes()
require.NoError(t, err)
assert.Equal(t, lo0.Index, routeInterfaceIn(msgs, netip.MustParsePrefix(cidr)))
}
// Only naming a different utun asks for a restart. Empty, plain utun and names that aren't utunN take whichever is free
func TestTunDevNeedsRestart(t *testing.T) {
for name, want := range map[string]bool{
"": false,
"utun": false,
"nope": false,
"utun7": false,
"utun3": true,
} {
assert.Equal(t, want, tunDevNeedsRestart(name, "utun7"), name)
}
}
+121
View File
@@ -0,0 +1,121 @@
//go:build !ios && !e2e_testing
package overlay
import (
"bytes"
"os"
"syscall"
"testing"
"github.com/stretchr/testify/require"
)
func TestTun_WriteFraming(t *testing.T) {
r, w, err := os.Pipe()
require.NoError(t, err)
defer r.Close()
defer w.Close()
tn := &tun{f: w}
for _, tc := range []struct {
name string
pkt []byte
af byte
}{
// Literal families rather than syscall.AF_*, so a wrong constant in Write can't also fix the expectation.
{"ipv4", []byte{0x45, 1, 2, 3, 4, 5}, 2},
{"ipv6", []byte{0x60, 1, 2, 3, 4, 5}, 30},
} {
t.Run(tc.name, func(t *testing.T) {
n, err := tn.Write(tc.pkt)
require.NoError(t, err)
require.Equal(t, len(tc.pkt), n)
got := make([]byte, 64)
n, err = r.Read(got)
require.NoError(t, err)
require.Equal(t, append([]byte{0, 0, 0, tc.af}, tc.pkt...), got[:n])
})
}
_, err = tn.Write([]byte{0x10, 1, 2, 3})
require.Error(t, err)
_, err = tn.Write([]byte{})
require.ErrorIs(t, err, syscall.EIO)
// The rejected writes must not have put anything on the fd, so the next read sees only this packet.
pkt := []byte{0x45, 9, 9}
_, err = tn.Write(pkt)
require.NoError(t, err)
got := make([]byte, 64)
n, err := r.Read(got)
require.NoError(t, err)
require.Equal(t, append([]byte{0, 0, 0, 2}, pkt...), got[:n])
}
func TestTun_ReadStripsHeader(t *testing.T) {
r, w, err := os.Pipe()
require.NoError(t, err)
defer r.Close()
defer w.Close()
tn := &tun{f: r}
pkt := []byte{0x45, 1, 2, 3, 4, 5}
_, err = w.Write(append([]byte{0, 0, 0, syscall.AF_INET}, pkt...))
require.NoError(t, err)
got := make([]byte, 64)
n, err := tn.Read(got)
require.NoError(t, err)
require.Equal(t, pkt, got[:n])
}
func TestTun_ReadShort(t *testing.T) {
r, w, err := os.Pipe()
require.NoError(t, err)
defer r.Close()
defer w.Close()
tn := &tun{f: r}
// Fewer bytes than the protocol header is not a packet.
_, err = w.Write([]byte{0, 0})
require.NoError(t, err)
n, err := tn.Read(make([]byte, 64))
require.NoError(t, err)
require.Zero(t, n)
// The next packet reads cleanly only if the short read consumed exactly what was written.
pkt := []byte{0x45, 1, 2, 3}
_, err = w.Write(append([]byte{0, 0, 0, syscall.AF_INET}, pkt...))
require.NoError(t, err)
got := make([]byte, 64)
n, err = tn.Read(got)
require.NoError(t, err)
require.Equal(t, pkt, got[:n])
}
// A heap allocation per packet rarely shows in a short throughput run, only across GC pauses, so
// this pins Read and Write to zero allocations directly.
func TestTun_ReadWriteDoNotAllocate(t *testing.T) {
r, w, err := os.Pipe()
require.NoError(t, err)
defer r.Close()
defer w.Close()
tw := &tun{f: w}
tr := &tun{f: r}
pkts := [][]byte{bytes.Repeat([]byte{0x45}, 1300), bytes.Repeat([]byte{0x60}, 1300)}
buf := make([]byte, 1500)
i := 0
allocs := testing.AllocsPerRun(1000, func() {
i++
if _, err := tw.Write(pkts[i%len(pkts)]); err != nil {
t.Fatal(err)
}
if _, err := tr.Read(buf); err != nil {
t.Fatal(err)
}
})
require.Zero(t, allocs)
}
+23 -13
View File
@@ -370,7 +370,7 @@ func newTun(c *config.C, l *slog.Logger, vpnNetworks []netip.Prefix, _ bool) (*t
t := &tun{
Device: deviceName,
vpnNetworks: vpnNetworks,
MTU: c.GetInt("tun.mtu", DefaultMTU),
MTU: getMTU(l, c),
l: l,
fd: fd,
shutdownR: shutdownR,
@@ -385,16 +385,18 @@ func newTun(c *config.C, l *slog.Logger, vpnNetworks []netip.Prefix, _ bool) (*t
},
}
err = t.reload(c, true)
err = t.reload(c, true, nil)
if err != nil {
return nil, err
}
closeOnErr = false
c.RegisterReloadCallback(func(c *config.C) {
err := t.reload(c, false)
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
r.RestartRequiredIfChanged("tun.dev", "tun.mtu")
err := t.reload(c, false, r)
if err != nil {
util.LogWithContextIfNeeded("failed to reload tun device", err, t.l)
r.Failed()
}
})
@@ -493,7 +495,7 @@ func (t *tun) Activate() error {
}
}
return t.addRoutes(false)
return t.addRoutes(false, nil)
}
func (t *tun) setMTU() error {
@@ -509,8 +511,8 @@ func (t *tun) setMTU() error {
return err
}
func (t *tun) reload(c *config.C, initial bool) error {
change, routes, err := getAllRoutesFromConfig(c, t.vpnNetworks, initial)
func (t *tun) reload(c *config.C, initial bool, r *config.Reload) error {
change, routes, err := getAllRoutesFromConfig(t.l, c, t.vpnNetworks, initial)
if err != nil {
return err
}
@@ -530,16 +532,18 @@ func (t *tun) reload(c *config.C, initial bool) error {
if !initial {
// Remove first, if the system removes a wanted route hopefully it will be re-added next
err := t.removeRoutes(findRemovedRoutes(routes, *oldRoutes))
err := t.removeRoutes(findRemovedRoutes(routes, *oldRoutes), r)
if err != nil {
util.LogWithContextIfNeeded("Failed to remove routes", err, t.l)
r.Failed()
}
// Ensure any routes we actually want are installed
err = t.addRoutes(true)
err = t.addRoutes(true, r)
if err != nil {
// Catch any stray logs
util.LogWithContextIfNeeded("Failed to add routes", err, t.l)
r.Failed()
}
}
@@ -563,7 +567,7 @@ func (t *tun) Queues(int) ([]tio.Queue, error) {
return []tio.Queue{tio.NewSingleQueue(t, defaultBatchBufSize)}, nil
}
func (t *tun) addRoutes(logErrors bool) error {
func (t *tun) addRoutes(logErrors bool, rl *config.Reload) error {
routes := *t.Routes.Load()
for _, r := range routes {
if len(r.Via) == 0 || !r.Install {
@@ -576,6 +580,7 @@ func (t *tun) addRoutes(logErrors bool) error {
retErr := util.NewContextualError("Failed to add route", map[string]any{"route": r}, err)
if logErrors {
retErr.Log(t.l)
rl.Failed()
} else {
return retErr
}
@@ -587,15 +592,20 @@ func (t *tun) addRoutes(logErrors bool) error {
return nil
}
func (t *tun) removeRoutes(routes []Route) error {
func (t *tun) removeRoutes(routes []Route, rl *config.Reload) error {
for _, r := range routes {
if !r.Install {
if len(r.Via) == 0 || !r.Install {
// addRoutes never installs these
continue
}
err := delRoute(r.Cidr, t.linkAddr)
if err != nil {
if errors.Is(err, unix.ESRCH) {
// Already gone, which is what we wanted
t.l.Info("Route was already removed", "route", r)
} else if err != nil {
t.l.Error("Failed to remove route", "error", err, "route", r)
rl.Failed()
} else {
t.l.Info("Removed route", "route", r)
}
+6 -5
View File
@@ -47,16 +47,17 @@ func newTunFromFd(c *config.C, l *slog.Logger, deviceFd int, vpnNetworks []netip
l: l,
}
err := t.reload(c, true)
err := t.reload(c, true, nil)
if err != nil {
_ = file.Close()
return nil, err
}
c.RegisterReloadCallback(func(c *config.C) {
err := t.reload(c, false)
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
err := t.reload(c, false, r)
if err != nil {
util.LogWithContextIfNeeded("failed to reload tun device", err, t.l)
r.Failed()
}
})
@@ -67,8 +68,8 @@ func (t *tun) Activate() error {
return nil
}
func (t *tun) reload(c *config.C, initial bool) error {
change, routes, err := getAllRoutesFromConfig(c, t.vpnNetworks, initial)
func (t *tun) reload(c *config.C, initial bool, r *config.Reload) error {
change, routes, err := getAllRoutesFromConfig(t.l, c, t.vpnNetworks, initial)
if err != nil {
return err
}
+41 -18
View File
@@ -185,6 +185,10 @@ func newTun(c *config.C, l *slog.Logger, vpnNetworks []netip.Prefix, multiqueue
}
t.Device = name
// Only read when we create the device, an fd handed to us already has its name and offloads
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
r.RestartRequiredIfChanged("tun.dev", "tun.use_offloads")
})
return t, nil
}
@@ -221,30 +225,32 @@ func newTunGeneric(c *config.C, l *slog.Logger, fd int, vnetHdr bool, offloadFla
vnetHdr: vnetHdr,
offloadFlags: offloadFlags,
vpnNetworks: vpnNetworks,
TXQueueLen: c.GetInt("tun.tx_queue", 500),
TXQueueLen: c.GetInt("tun.tx_queue", 5000),
useSystemRoutes: c.GetBool("tun.use_system_route_table", false),
useSystemRoutesBufferSize: c.GetInt("tun.use_system_route_table_buffer_size", 0),
routesFromSystem: map[netip.Prefix]routing.Gateways{},
l: l,
}
if err = t.reload(c, true); err != nil {
if err = t.reload(c, true, nil); err != nil {
_ = t.Close()
return nil, err
}
c.RegisterReloadCallback(func(c *config.C) {
err := t.reload(c, false)
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
r.RestartRequiredIfChanged("tun.tx_queue", "tun.use_system_route_table", "tun.use_system_route_table_buffer_size")
err := t.reload(c, false, r)
if err != nil {
util.LogWithContextIfNeeded("failed to reload tun device", err, t.l)
r.Failed()
}
})
return t, nil
}
func (t *tun) reload(c *config.C, initial bool) error {
routeChange, routes, err := getAllRoutesFromConfig(c, t.vpnNetworks, initial)
func (t *tun) reload(c *config.C, initial bool, r *config.Reload) error {
routeChange, routes, err := getAllRoutesFromConfig(t.l, c, t.vpnNetworks, initial)
if err != nil {
return err
}
@@ -253,6 +259,15 @@ func (t *tun) reload(c *config.C, initial bool) error {
return nil
}
// Only the mtu changed. The routes are the same, parse them again so the ones without their own mtu pick up the
// new default. Without this an mtu change would remove them all
if !routeChange {
_, routes, err = getAllRoutesFromConfig(t.l, c, t.vpnNetworks, true)
if err != nil {
return err
}
}
routeTree, err := makeRouteTree(t.l, routes, true)
if err != nil {
return err
@@ -267,14 +282,14 @@ func (t *tun) reload(c *config.C, initial bool) error {
oldDefaultMTU := t.DefaultMTU
oldMaxMTU := t.MaxMTU
newDefaultMTU := c.GetInt("tun.mtu", DefaultMTU)
newDefaultMTU := getMTU(t.l, c)
newMaxMTU := newDefaultMTU
for i, r := range routes {
if r.MTU == 0 {
routes[i].MTU = newDefaultMTU
}
if r.MTU > t.MaxMTU {
if r.MTU > newMaxMTU {
newMaxMTU = r.MTU
}
}
@@ -288,7 +303,7 @@ func (t *tun) reload(c *config.C, initial bool) error {
if !initial {
if oldMaxMTU != newMaxMTU {
t.setMTU()
t.setMTU(r)
t.l.Info("Set max MTU", "mtu", t.MaxMTU, "oldMTU", oldMaxMTU)
}
@@ -296,7 +311,8 @@ func (t *tun) reload(c *config.C, initial bool) error {
for i := range t.vpnNetworks {
err := t.setDefaultRoute(t.vpnNetworks[i])
if err != nil {
t.l.Warn(err.Error())
t.l.Error(err.Error())
r.Failed()
} else {
t.l.Info("Set default MTU", "mtu", t.DefaultMTU, "oldMTU", oldDefaultMTU)
}
@@ -304,13 +320,14 @@ func (t *tun) reload(c *config.C, initial bool) error {
}
// Remove first, if the system removes a wanted route hopefully it will be re-added next
t.removeRoutes(findRemovedRoutes(routes, *oldRoutes))
t.removeRoutes(findRemovedRoutes(routes, *oldRoutes), r)
// Ensure any routes we actually want are installed
err = t.addRoutes(true)
err = t.addRoutes(true, r)
if err != nil {
// This should never be called since addRoutes should log its own errors in a reload condition
util.LogWithContextIfNeeded("Failed to refresh routes", err, t.l)
r.Failed()
}
}
@@ -465,7 +482,7 @@ func (t *tun) Activate() error {
t.deviceIndex = link.Attrs().Index
// Setup our default MTU
t.setMTU()
t.setMTU(nil)
// Set the transmit queue length
ifrq := ifreqQLEN{Name: devName, Value: int32(t.TXQueueLen)}
@@ -497,7 +514,7 @@ func (t *tun) Activate() error {
}
// Set the routes
if err = t.addRoutes(false); err != nil {
if err = t.addRoutes(false, nil); err != nil {
return err
}
@@ -510,12 +527,13 @@ func (t *tun) Activate() error {
return nil
}
func (t *tun) setMTU() {
func (t *tun) setMTU(r *config.Reload) {
// Set the MTU on the device
ifm := ifreqMTU{Name: t.deviceBytes(), MTU: int32(t.MaxMTU)}
if err := ioctl(t.ioctlFd, unix.SIOCSIFMTU, uintptr(unsafe.Pointer(&ifm))); err != nil {
// This is currently a non fatal condition because the route table must have the MTU set appropriately as well
t.l.Error("Failed to set tun mtu", "error", err)
r.Failed()
}
}
@@ -568,7 +586,7 @@ func (t *tun) setDefaultRoute(cidr netip.Prefix) error {
return nil
}
func (t *tun) addRoutes(logErrors bool) error {
func (t *tun) addRoutes(logErrors bool, rl *config.Reload) error {
// Path routes
routes := *t.Routes.Load()
for _, r := range routes {
@@ -598,6 +616,7 @@ func (t *tun) addRoutes(logErrors bool) error {
retErr := util.NewContextualError("Failed to add route", map[string]any{"route": r}, err)
if logErrors {
retErr.Log(t.l)
rl.Failed()
} else {
return retErr
}
@@ -609,7 +628,7 @@ func (t *tun) addRoutes(logErrors bool) error {
return nil
}
func (t *tun) removeRoutes(routes []Route) {
func (t *tun) removeRoutes(routes []Route, rl *config.Reload) {
for _, r := range routes {
if !r.Install {
continue
@@ -633,8 +652,12 @@ func (t *tun) removeRoutes(routes []Route) {
}
err := netlink.RouteDel(&nr)
if err != nil {
if errors.Is(err, unix.ESRCH) {
// Already gone, which is what we wanted
t.l.Info("Route was already removed", "route", r)
} else if err != nil {
t.l.Error("Failed to remove route", "error", err, "route", r)
rl.Failed()
} else {
t.l.Info("Removed route", "route", r)
}
+27 -27
View File
@@ -8,7 +8,6 @@ import (
"log/slog"
"net/netip"
"os"
"regexp"
"sync/atomic"
"syscall"
"unsafe"
@@ -67,24 +66,12 @@ type tun struct {
fd int
}
var deviceNameRE = regexp.MustCompile(`^tun[0-9]+$`)
func newTunFromFd(_ *config.C, _ *slog.Logger, _ int, _ []netip.Prefix) (*tun, error) {
return nil, fmt.Errorf("newTunFromFd not supported in NetBSD")
}
func newTun(c *config.C, l *slog.Logger, vpnNetworks []netip.Prefix, _ bool) (*tun, error) {
// Try to open tun device
var err error
deviceName := c.GetString("tun.dev", "")
if deviceName == "" {
return nil, fmt.Errorf("a device name in the format of /dev/tunN must be specified")
}
if !deviceNameRE.MatchString(deviceName) {
return nil, fmt.Errorf("a device name in the format of /dev/tunN must be specified")
}
fd, err := unix.Open("/dev/"+deviceName, os.O_RDWR, 0)
fd, deviceName, err := openTun(l, c.GetString("tun.dev", ""))
if err != nil {
return nil, err
}
@@ -99,19 +86,24 @@ func newTun(c *config.C, l *slog.Logger, vpnNetworks []netip.Prefix, _ bool) (*t
fd: fd,
Device: deviceName,
vpnNetworks: vpnNetworks,
MTU: c.GetInt("tun.mtu", DefaultMTU),
MTU: getMTU(l, c),
l: l,
}
err = t.reload(c, true)
err = t.reload(c, true, nil)
if err != nil {
return nil, err
}
c.RegisterReloadCallback(func(c *config.C) {
err := t.reload(c, false)
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
r.RestartRequiredIfChanged("tun.mtu")
if tunDevNeedsRestart(c.GetString("tun.dev", ""), t.Device) {
r.RestartRequiredIfChanged("tun.dev")
}
err := t.reload(c, false, r)
if err != nil {
util.LogWithContextIfNeeded("failed to reload tun device", err, t.l)
r.Failed()
}
})
@@ -324,7 +316,7 @@ func (t *tun) Activate() error {
}
}
return t.addRoutes(false)
return t.addRoutes(false, nil)
}
func (t *tun) doIoctlByName(ctl uintptr, value uint32) error {
@@ -339,8 +331,8 @@ func (t *tun) doIoctlByName(ctl uintptr, value uint32) error {
return err
}
func (t *tun) reload(c *config.C, initial bool) error {
change, routes, err := getAllRoutesFromConfig(c, t.vpnNetworks, initial)
func (t *tun) reload(c *config.C, initial bool, r *config.Reload) error {
change, routes, err := getAllRoutesFromConfig(t.l, c, t.vpnNetworks, initial)
if err != nil {
return err
}
@@ -360,16 +352,18 @@ func (t *tun) reload(c *config.C, initial bool) error {
if !initial {
// Remove first, if the system removes a wanted route hopefully it will be re-added next
err := t.removeRoutes(findRemovedRoutes(routes, *oldRoutes))
err := t.removeRoutes(findRemovedRoutes(routes, *oldRoutes), r)
if err != nil {
util.LogWithContextIfNeeded("Failed to remove routes", err, t.l)
r.Failed()
}
// Ensure any routes we actually want are installed
err = t.addRoutes(true)
err = t.addRoutes(true, r)
if err != nil {
// Catch any stray logs
util.LogWithContextIfNeeded("Failed to add routes", err, t.l)
r.Failed()
}
}
@@ -393,7 +387,7 @@ func (t *tun) Queues(int) ([]tio.Queue, error) {
return []tio.Queue{tio.NewSingleQueue(t, defaultBatchBufSize)}, nil
}
func (t *tun) addRoutes(logErrors bool) error {
func (t *tun) addRoutes(logErrors bool, rl *config.Reload) error {
routes := *t.Routes.Load()
for _, r := range routes {
@@ -407,6 +401,7 @@ func (t *tun) addRoutes(logErrors bool) error {
retErr := util.NewContextualError("Failed to add route", map[string]any{"route": r}, err)
if logErrors {
retErr.Log(t.l)
rl.Failed()
} else {
return retErr
}
@@ -418,15 +413,20 @@ func (t *tun) addRoutes(logErrors bool) error {
return nil
}
func (t *tun) removeRoutes(routes []Route) error {
func (t *tun) removeRoutes(routes []Route, rl *config.Reload) error {
for _, r := range routes {
if !r.Install {
if len(r.Via) == 0 || !r.Install {
// addRoutes never installs these
continue
}
err := delRoute(r.Cidr, t.vpnNetworks)
if err != nil {
if errors.Is(err, unix.ESRCH) {
// Already gone, which is what we wanted
t.l.Info("Route was already removed", "route", r)
} else if err != nil {
t.l.Error("Failed to remove route", "error", err, "route", r)
rl.Failed()
} else {
t.l.Info("Removed route", "route", r)
}
+27 -27
View File
@@ -8,7 +8,6 @@ import (
"log/slog"
"net/netip"
"os"
"regexp"
"sync/atomic"
"syscall"
"unsafe"
@@ -58,24 +57,12 @@ type tun struct {
fd int
}
var deviceNameRE = regexp.MustCompile(`^tun[0-9]+$`)
func newTunFromFd(_ *config.C, _ *slog.Logger, _ int, _ []netip.Prefix) (*tun, error) {
return nil, fmt.Errorf("newTunFromFd not supported in openbsd")
}
func newTun(c *config.C, l *slog.Logger, vpnNetworks []netip.Prefix, _ bool) (*tun, error) {
// Try to open tun device
var err error
deviceName := c.GetString("tun.dev", "")
if deviceName == "" {
return nil, fmt.Errorf("a device name in the format of /dev/tunN must be specified")
}
if !deviceNameRE.MatchString(deviceName) {
return nil, fmt.Errorf("a device name in the format of /dev/tunN must be specified")
}
fd, err := unix.Open("/dev/"+deviceName, os.O_RDWR, 0)
fd, deviceName, err := openTun(l, c.GetString("tun.dev", ""))
if err != nil {
return nil, err
}
@@ -90,19 +77,24 @@ func newTun(c *config.C, l *slog.Logger, vpnNetworks []netip.Prefix, _ bool) (*t
fd: fd,
Device: deviceName,
vpnNetworks: vpnNetworks,
MTU: c.GetInt("tun.mtu", DefaultMTU),
MTU: getMTU(l, c),
l: l,
}
err = t.reload(c, true)
err = t.reload(c, true, nil)
if err != nil {
return nil, err
}
c.RegisterReloadCallback(func(c *config.C) {
err := t.reload(c, false)
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
r.RestartRequiredIfChanged("tun.mtu")
if tunDevNeedsRestart(c.GetString("tun.dev", ""), t.Device) {
r.RestartRequiredIfChanged("tun.dev")
}
err := t.reload(c, false, r)
if err != nil {
util.LogWithContextIfNeeded("failed to reload tun device", err, t.l)
r.Failed()
}
})
@@ -303,7 +295,7 @@ func (t *tun) Activate() error {
}
}
return t.addRoutes(false)
return t.addRoutes(false, nil)
}
func (t *tun) doIoctlByName(ctl uintptr, value uint32) error {
@@ -318,8 +310,8 @@ func (t *tun) doIoctlByName(ctl uintptr, value uint32) error {
return err
}
func (t *tun) reload(c *config.C, initial bool) error {
change, routes, err := getAllRoutesFromConfig(c, t.vpnNetworks, initial)
func (t *tun) reload(c *config.C, initial bool, r *config.Reload) error {
change, routes, err := getAllRoutesFromConfig(t.l, c, t.vpnNetworks, initial)
if err != nil {
return err
}
@@ -339,16 +331,18 @@ func (t *tun) reload(c *config.C, initial bool) error {
if !initial {
// Remove first, if the system removes a wanted route hopefully it will be re-added next
err := t.removeRoutes(findRemovedRoutes(routes, *oldRoutes))
err := t.removeRoutes(findRemovedRoutes(routes, *oldRoutes), r)
if err != nil {
util.LogWithContextIfNeeded("Failed to remove routes", err, t.l)
r.Failed()
}
// Ensure any routes we actually want are installed
err = t.addRoutes(true)
err = t.addRoutes(true, r)
if err != nil {
// Catch any stray logs
util.LogWithContextIfNeeded("Failed to add routes", err, t.l)
r.Failed()
}
}
@@ -372,7 +366,7 @@ func (t *tun) Queues(int) ([]tio.Queue, error) {
return []tio.Queue{tio.NewSingleQueue(t, defaultBatchBufSize)}, nil
}
func (t *tun) addRoutes(logErrors bool) error {
func (t *tun) addRoutes(logErrors bool, rl *config.Reload) error {
routes := *t.Routes.Load()
for _, r := range routes {
@@ -386,6 +380,7 @@ func (t *tun) addRoutes(logErrors bool) error {
retErr := util.NewContextualError("Failed to add route", map[string]any{"route": r}, err)
if logErrors {
retErr.Log(t.l)
rl.Failed()
} else {
return retErr
}
@@ -397,15 +392,20 @@ func (t *tun) addRoutes(logErrors bool) error {
return nil
}
func (t *tun) removeRoutes(routes []Route) error {
func (t *tun) removeRoutes(routes []Route, rl *config.Reload) error {
for _, r := range routes {
if !r.Install {
if len(r.Via) == 0 || !r.Install {
// addRoutes never installs these
continue
}
err := delRoute(r.Cidr, t.vpnNetworks)
if err != nil {
if errors.Is(err, unix.ESRCH) {
// Already gone, which is what we wanted
t.l.Info("Route was already removed", "route", r)
} else if err != nil {
t.l.Error("Failed to remove route", "error", err, "route", r)
rl.Failed()
} else {
t.l.Info("Removed route", "route", r)
}
+107
View File
@@ -0,0 +1,107 @@
//go:build !e2e_testing
package overlay
import (
"net/netip"
"os"
"testing"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/routing"
"github.com/slackhq/nebula/test"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
// Only tun.mtu changes. The system calls fail without a real device, what matters is the routes nebula keeps
func TestReloadMTUOnlyKeepsRoutes(t *testing.T) {
routes := " unsafe_routes:\n - route: 192.168.100.0/24\n via: 10.1.0.1\n"
c := config.NewC(test.NewLogger())
require.NoError(t, c.LoadString("tun:\n mtu: 1300\n"+routes))
tn := &tun{
l: test.NewLogger(),
vpnNetworks: []netip.Prefix{netip.MustParsePrefix("10.1.0.2/24")},
ioctlFd: ^uintptr(0),
routesFromSystem: map[netip.Prefix]routing.Gateways{},
}
require.NoError(t, tn.reload(c, true, nil))
require.NoError(t, c.ReloadConfigString("tun:\n mtu: 1400\n"+routes).Err)
_ = tn.reload(c, false, nil)
got := *tn.Routes.Load()
require.Len(t, got, 1)
assert.Equal(t, netip.MustParsePrefix("192.168.100.0/24"), got[0].Cidr)
assert.Equal(t, 1400, got[0].MTU)
assert.Equal(t, 1400, tn.DefaultMTU)
}
// The device mtu has to fit the largest route mtu, wherever it sits in the list
func TestReloadMaxMTUIsTheLargestRoute(t *testing.T) {
c := config.NewC(test.NewLogger())
require.NoError(t, c.LoadString("tun:\n mtu: 1300\n unsafe_routes:\n"+
" - route: 192.168.100.0/24\n via: 10.1.0.1\n mtu: 1500\n"+
" - route: 192.168.101.0/24\n via: 10.1.0.1\n mtu: 1400\n"))
tn := &tun{
l: test.NewLogger(),
vpnNetworks: []netip.Prefix{netip.MustParsePrefix("10.1.0.2/24")},
ioctlFd: ^uintptr(0),
routesFromSystem: map[netip.Prefix]routing.Gateways{},
}
require.NoError(t, tn.reload(c, true, nil))
assert.Equal(t, 1500, tn.MaxMTU)
assert.Equal(t, 1300, tn.DefaultMTU)
}
// There is no device behind this tun, so adding a route always fails. Removing one is refused without root, as root the
// kernel answers that the route isn't there, which is what we wanted
func TestReloadRouteFailures(t *testing.T) {
route := "tun:\n mtu: 1300\n unsafe_routes:\n - route: 192.168.100.0/24\n via: 10.1.0.1\n"
cases := []struct {
name string
from, to string
wantFails bool
}{
{"added", "tun:\n mtu: 1300\n", route, true},
{"removed", route, "tun:\n mtu: 1300\n", os.Geteuid() != 0},
{"unchanged", route, route + "punchy:\n punch: true\n", false},
}
for _, tc := range cases {
t.Run(tc.name, func(t *testing.T) {
c := config.NewC(test.NewLogger())
require.NoError(t, c.LoadString(tc.from))
tn := &tun{
l: test.NewLogger(),
vpnNetworks: []netip.Prefix{netip.MustParsePrefix("10.1.0.2/24")},
ioctlFd: ^uintptr(0),
routesFromSystem: map[netip.Prefix]routing.Gateways{},
}
require.NoError(t, tn.reload(c, true, nil))
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
require.NoError(t, tn.reload(c, false, r))
})
res := c.ReloadConfigString(tc.to)
require.NoError(t, res.Err)
assert.Equal(t, tc.wantFails, res.Failed)
})
}
}
// Keys only read when the device is created need a restart to apply
func TestTunReloadRestartKeys(t *testing.T) {
c := config.NewC(test.NewLogger())
require.NoError(t, c.LoadString("tun:\n dev: nebrk0\n"))
tn, err := NewDeviceFromConfig(c, test.NewLogger(), []netip.Prefix{netip.MustParsePrefix("10.1.0.2/24")}, 1)
if err != nil {
t.Skipf("can't create a tun device here: %v", err)
}
t.Cleanup(func() { _ = tn.Close() })
res := c.ReloadConfigString("tun:\n disabled: true\n dev: nebrk1\n tx_queue: 100\n")
require.NoError(t, res.Err)
assert.ElementsMatch(t, []string{"tun.disabled", "tun.dev", "tun.tx_queue"}, res.RestartKeys)
}
+24
View File
@@ -0,0 +1,24 @@
package overlay
import (
"net/netip"
"testing"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/test"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
// tun.disabled and the disabled tun's settings are only read when the device is made
func TestDisabledTunRestartKeys(t *testing.T) {
c := config.NewC(test.NewLogger())
require.NoError(t, c.LoadString("tun:\n disabled: true\n"))
d, err := NewDeviceFromConfig(c, test.NewLogger(), []netip.Prefix{netip.MustParsePrefix("10.1.0.2/24")}, 1)
require.NoError(t, err)
t.Cleanup(func() { _ = d.Close() })
res := c.ReloadConfigString("tun:\n disabled: false\n tx_queue: 100\n")
require.NoError(t, res.Err)
assert.ElementsMatch(t, []string{"tun.disabled", "tun.tx_queue"}, res.RestartKeys)
}
+1 -1
View File
@@ -31,7 +31,7 @@ type TestTun struct {
}
func newTun(c *config.C, l *slog.Logger, vpnNetworks []netip.Prefix, _ bool) (*TestTun, error) {
_, routes, err := getAllRoutesFromConfig(c, vpnNetworks, true)
_, routes, err := getAllRoutesFromConfig(l, c, vpnNetworks, true)
if err != nil {
return nil, err
}
+106 -14
View File
@@ -4,6 +4,7 @@ package overlay
import (
"crypto"
"errors"
"fmt"
"log/slog"
"net/netip"
@@ -75,7 +76,7 @@ func newTun(c *config.C, l *slog.Logger, vpnNetworks []netip.Prefix, _ bool) (*w
t := &winTun{
Device: deviceName,
vpnNetworks: vpnNetworks,
MTU: c.GetInt("tun.mtu", DefaultMTU),
MTU: getMTU(l, c),
guid: *guid,
networkCategory: cat,
setCategory: setCat,
@@ -83,7 +84,7 @@ func newTun(c *config.C, l *slog.Logger, vpnNetworks []netip.Prefix, _ bool) (*w
l: l,
}
err = t.reload(c, true)
err = t.reload(c, true, nil)
if err != nil {
return nil, err
}
@@ -104,18 +105,32 @@ func newTun(c *config.C, l *slog.Logger, vpnNetworks []netip.Prefix, _ bool) (*w
}
t.tun = tunDevice.(*wintun.NativeTun)
c.RegisterReloadCallback(func(c *config.C) {
err := t.reload(c, false)
// Disable registration before Activate assigns addresses and triggers it.
disableDNSRegistration(l, *guid)
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
r.RestartRequiredIfChanged("tun.dev", "tun.mtu")
// Compare with the running category, not the last config. A restart with one we can't parse would fail, so say
// that now rather than ask for one, and going back to the running one after a bad reload needs no restart
if cat, setCat, err := parseNetworkCategory(c.GetString("tun.network_category", "private")); err != nil {
t.l.Error("Invalid tun.network_category, a restart with it would fail", "error", err)
r.Failed()
} else if cat != t.networkCategory || setCat != t.setCategory {
r.RestartRequiredIfChanged("tun.network_category")
}
r.RestartRequiredIfChanged(interfaceBypassKeys...)
err := t.reload(c, false, r)
if err != nil {
util.LogWithContextIfNeeded("failed to reload tun device", err, t.l)
r.Failed()
}
})
return t, nil
}
func (t *winTun) reload(c *config.C, initial bool) error {
change, routes, err := getAllRoutesFromConfig(c, t.vpnNetworks, initial)
func (t *winTun) reload(c *config.C, initial bool, r *config.Reload) error {
change, routes, err := getAllRoutesFromConfig(t.l, c, t.vpnNetworks, initial)
if err != nil {
return err
}
@@ -135,16 +150,18 @@ func (t *winTun) reload(c *config.C, initial bool) error {
if !initial {
// Remove first, if the system removes a wanted route hopefully it will be re-added next
err := t.removeRoutes(findRemovedRoutes(routes, *oldRoutes))
err := t.removeRoutes(findRemovedSystemRoutes(routes, *oldRoutes), r)
if err != nil {
util.LogWithContextIfNeeded("Failed to remove routes", err, t.l)
r.Failed()
}
// Ensure any routes we actually want are installed
err = t.addRoutes(true)
err = t.addRoutes(true, r)
if err != nil {
// Catch any stray logs
util.LogWithContextIfNeeded("Failed to add routes", err, t.l)
r.Failed()
}
}
@@ -159,7 +176,7 @@ func (t *winTun) Activate() error {
return fmt.Errorf("failed to set address: %w", err)
}
err = t.addRoutes(false)
err = t.addRoutes(false, nil)
if err != nil {
return err
}
@@ -178,33 +195,49 @@ func (t *winTun) Activate() error {
return nil
}
func (t *winTun) addRoutes(logErrors bool) error {
func (t *winTun) addRoutes(logErrors bool, rl *config.Reload) error {
luid := winipcfg.LUID(t.tun.LUID())
routes := *t.Routes.Load()
foundDefault4 := false
carriesV6 := slices.ContainsFunc(t.vpnNetworks, func(p netip.Prefix) bool { return p.Addr().Is6() })
for _, r := range routes {
// A CIDR listed twice routes by its last entry, install only that one or each reload would flip between them. The
// same rule as systemRoutes
last := make(map[netip.Prefix]int, len(routes))
for i, r := range routes {
last[r.Cidr] = i
}
for i, r := range routes {
if len(r.Via) == 0 || !r.Install {
// We don't allow route MTUs so only install routes with a via
continue
}
if last[r.Cidr] != i {
continue
}
// A v6 unsafe_route is legal under a v4-only cert; uninstalled ones put nothing on the adapter.
carriesV6 = carriesV6 || r.Cidr.Addr().Is6()
// Add our unsafe route as an on-link route to the nebula tun device.
err := luid.AddRoute(r.Cidr, unspecifiedNextHop(r.Cidr), uint32(r.Metric))
result, err := addRoute(luid, r)
if err != nil {
retErr := util.NewContextualError("Failed to add route", map[string]any{"route": r}, err)
if logErrors {
retErr.Log(t.l)
rl.Failed()
continue
} else {
return retErr
}
} else {
switch result {
case routeAdded:
t.l.Info("Added route", "route", r)
case routeMetricUpdated:
t.l.Info("Updated route metric", "route", r)
}
}
if !foundDefault4 {
@@ -255,7 +288,7 @@ func (t *winTun) setMTU(luid winipcfg.LUID, foundDefault4, carriesV6 bool) error
return nil
}
func (t *winTun) removeRoutes(routes []Route) error {
func (t *winTun) removeRoutes(routes []Route, rl *config.Reload) error {
luid := winipcfg.LUID(t.tun.LUID())
for _, r := range routes {
@@ -265,8 +298,12 @@ func (t *winTun) removeRoutes(routes []Route) error {
// See comment on luid.AddRoute
err := luid.DeleteRoute(r.Cidr, unspecifiedNextHop(r.Cidr))
if err != nil {
if errors.Is(err, windows.ERROR_NOT_FOUND) {
// Already gone, which is what we wanted
t.l.Info("Route was already removed", "route", r)
} else if err != nil {
t.l.Error("Failed to remove route", "error", err, "route", r)
rl.Failed()
} else {
t.l.Info("Removed route", "route", r)
}
@@ -316,6 +353,61 @@ func (t *winTun) Close() error {
return t.tun.Close()
}
// systemRoutes maps each CIDR to the route that puts it in the system table. A CIDR's last entry decides, including
// one that says not to install it
func systemRoutes(routes []Route) map[netip.Prefix]Route {
out := make(map[netip.Prefix]Route, len(routes))
for _, r := range routes {
if len(r.Via) > 0 && r.Install {
out[r.Cidr] = r
} else {
delete(out, r.Cidr)
}
}
return out
}
// findRemovedSystemRoutes is the routes to delete from the system table. Windows deletes by destination and next hop,
// so only a CIDR nothing installs anymore goes, a changed metric is set on the route in place by addRoute
func findRemovedSystemRoutes(newRoutes, oldRoutes []Route) []Route {
keep := systemRoutes(newRoutes)
var removed []Route
for _, r := range systemRoutes(oldRoutes) {
if _, ok := keep[r.Cidr]; !ok {
removed = append(removed, r)
}
}
return removed
}
type routeResult int
const (
routeAdded routeResult = iota
routeMetricUpdated
routeAlreadyInstalled
)
// addRoute installs r on the interface. A reload re-adds every route to catch any the system dropped, and Windows
// refuses one it has with the same destination and next hop whatever its metric. One that is there with another
// metric, ours from before a metric change or someone else's, is brought to ours in place
func addRoute(luid winipcfg.LUID, r Route) (routeResult, error) {
nextHop := unspecifiedNextHop(r.Cidr)
err := luid.AddRoute(r.Cidr, nextHop, uint32(r.Metric))
if !errors.Is(err, windows.ERROR_OBJECT_ALREADY_EXISTS) {
return routeAdded, err
}
row, err := luid.Route(r.Cidr, nextHop)
if err != nil {
return routeAlreadyInstalled, err
}
if row.Metric != uint32(r.Metric) {
row.Metric = uint32(r.Metric)
return routeMetricUpdated, row.Set()
}
return routeAlreadyInstalled, nil
}
func unspecifiedNextHop(p netip.Prefix) netip.Addr {
if p.Addr().Is4() {
return netip.IPv4Unspecified()
+209
View File
@@ -0,0 +1,209 @@
//go:build windows && !e2e_testing
package overlay
import (
"bytes"
"fmt"
"net/netip"
"strings"
"testing"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/test"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
"golang.org/x/sys/windows"
"golang.zx2c4.com/wireguard/windows/tunnel/winipcfg"
)
type routeTest struct {
c *config.C
tun *winTun
luid winipcfg.LUID
log *bytes.Buffer
}
// newRouteTest makes a wintun device of its own, before installing tun.unsafe_routes from cfg
func newRouteTest(t *testing.T, cfg string) *routeTest {
t.Helper()
if !windows.GetCurrentProcessToken().IsElevated() {
t.Skip("creating a wintun device needs an elevated process")
}
if err := checkWinTunExists(); err != nil {
t.Skipf("wintun.dll is not where nebula looks for it: %v", err)
}
rt := &routeTest{log: &bytes.Buffer{}}
l := test.NewLoggerWithOutput(rt.log)
rt.c = config.NewC(l)
require.NoError(t, rt.c.LoadString(cfg))
var err error
rt.tun, err = newTun(rt.c, l, []netip.Prefix{netip.MustParsePrefix("10.250.0.1/24")}, false)
require.NoError(t, err)
t.Cleanup(func() { _ = rt.tun.Close() })
rt.luid = winipcfg.LUID(rt.tun.tun.LUID())
return rt
}
func (rt *routeTest) metric(t *testing.T, cidr string) uint32 {
t.Helper()
p := netip.MustParsePrefix(cidr)
row, err := rt.luid.Route(p, unspecifiedNextHop(p))
require.NoError(t, err)
return row.Metric
}
// reload applies cfg and expects nothing in it to fail
func (rt *routeTest) reload(t *testing.T, cfg string) {
t.Helper()
res := rt.c.ReloadConfigString(cfg)
require.NoError(t, res.Err)
assert.False(t, res.Failed, "the reload reported a failure")
}
func unsafeRoutes(routes ...string) string {
var b strings.Builder
b.WriteString("tun:\n dev: nebroutetest\n unsafe_routes:\n")
// cidr[@metric][!], the ! for install: false
for _, r := range routes {
r, noInstall := strings.CutSuffix(r, "!")
cidr, metric, _ := strings.Cut(r, "@")
fmt.Fprintf(&b, " - route: %s\n via: 10.250.0.2\n", cidr)
if metric != "" {
fmt.Fprintf(&b, " metric: %s\n", metric)
}
if noInstall {
b.WriteString(" install: false\n")
}
}
return b.String()
}
// A reload that adds a route re-adds the ones already installed, Windows refusing those as already there isn't a failure
func TestWinTun_ReloadKeepsInstalledRoutes(t *testing.T) {
forFamilies(t, func(t *testing.T, a, b string) {
rt := newRouteTest(t, unsafeRoutes(a))
require.NoError(t, rt.tun.Activate())
rt.reload(t, unsafeRoutes(a, b))
assert.NotContains(t, rt.log.String(), "Failed to add route")
rt.metric(t, a)
rt.metric(t, b)
})
}
// Windows calls a route the same by destination and next hop, so a changed metric is set on the installed route in
// place rather than deleted and added, which would leave a moment with no route
func TestWinTun_ReloadMetricChangeApplies(t *testing.T) {
forFamilies(t, func(t *testing.T, a, b string) {
rt := newRouteTest(t, unsafeRoutes(a+"@100"))
require.NoError(t, rt.tun.Activate())
require.Equal(t, uint32(100), rt.metric(t, a))
rt.reload(t, unsafeRoutes(a+"@200"))
assert.NotContains(t, rt.log.String(), "Failed to add route")
assert.NotContains(t, rt.log.String(), "Removed route")
assert.Equal(t, uint32(200), rt.metric(t, a))
})
}
// A route someone else put on the device with another metric is brought to ours, at startup too
func TestWinTun_ForeignRouteGetsOurMetric(t *testing.T) {
forFamilies(t, func(t *testing.T, a, b string) {
rt := newRouteTest(t, unsafeRoutes(a+"@100"))
p := netip.MustParsePrefix(a)
require.NoError(t, rt.luid.AddRoute(p, unspecifiedNextHop(p), 50))
require.NoError(t, rt.tun.Activate())
assert.Equal(t, uint32(100), rt.metric(t, a))
assert.Contains(t, rt.log.String(), "level=INFO msg=\"Updated route metric\"")
})
}
// A CIDR listed twice is installed once, by its last entry, and reloads leave it there instead of flipping between the
// two metrics
func TestWinTun_DuplicateCidrInstallsTheLastEntry(t *testing.T) {
forFamilies(t, func(t *testing.T, a, b string) {
rt := newRouteTest(t, unsafeRoutes(a+"@100", a+"@200"))
require.NoError(t, rt.tun.Activate())
assert.Equal(t, uint32(200), rt.metric(t, a))
rt.reload(t, unsafeRoutes(a+"@100", a+"@200", b))
assert.Equal(t, uint32(200), rt.metric(t, a))
assert.NotContains(t, rt.log.String(), "Updated route metric")
assert.NotContains(t, rt.log.String(), "Failed to add route")
})
}
// Dropping the duplicate that wasn't installed leaves the installed route alone. Windows deletes by destination and
// next hop, deleting the dropped one would take the live route with it
func TestWinTun_DroppingAnUninstalledDuplicateKeepsTheRoute(t *testing.T) {
forFamilies(t, func(t *testing.T, a, b string) {
rt := newRouteTest(t, unsafeRoutes(a+"@100", a+"@200"))
require.NoError(t, rt.tun.Activate())
rt.reload(t, unsafeRoutes(a+"@200"))
assert.NotContains(t, rt.log.String(), "Removed route")
assert.Equal(t, uint32(200), rt.metric(t, a))
})
}
// A reload ends where a fresh start with the same config would, a CIDR whose last entry says not to install it is
// removed
func TestWinTun_ReloadMatchesAFreshStart(t *testing.T) {
forFamilies(t, func(t *testing.T, a, b string) {
cidr := netip.MustParsePrefix(a)
routes := unsafeRoutes(a+"@100", a+"@100!")
fresh := newRouteTest(t, routes)
require.NoError(t, fresh.tun.Activate())
_, err := fresh.luid.Route(cidr, unspecifiedNextHop(cidr))
require.Error(t, err, "a fresh start installed it")
require.NoError(t, fresh.tun.Close())
rt := newRouteTest(t, unsafeRoutes(a+"@100"))
require.NoError(t, rt.tun.Activate())
rt.reload(t, routes)
_, err = rt.luid.Route(cidr, unspecifiedNextHop(cidr))
assert.Error(t, err, "the reload left it installed")
})
}
// Removing a route that is already gone is what we wanted, not a failure
func TestWinTun_ReloadRouteAlreadyGone(t *testing.T) {
forFamilies(t, func(t *testing.T, a, b string) {
rt := newRouteTest(t, unsafeRoutes(a, b))
require.NoError(t, rt.tun.Activate())
p := netip.MustParsePrefix(b)
require.NoError(t, rt.luid.DeleteRoute(p, unspecifiedNextHop(p)))
rt.reload(t, unsafeRoutes(a))
})
}
// A network category we can't parse would fail the restart, so it fails the reload instead of asking for one
func TestWinTun_ReloadInvalidNetworkCategoryFails(t *testing.T) {
rt := newRouteTest(t, unsafeRoutes())
for range 2 {
res := rt.c.ReloadConfigString(unsafeRoutes() + " network_category: nope\n")
require.NoError(t, res.Err)
assert.True(t, res.Failed, "a bad category that didn't change since the last reload is still bad")
assert.False(t, res.RestartRequired)
}
// Fixing it back to the running category needs nothing
res := rt.c.ReloadConfigString(unsafeRoutes() + " network_category: private\n")
require.NoError(t, res.Err)
assert.False(t, res.Failed)
assert.False(t, res.RestartRequired)
res = rt.c.ReloadConfigString(unsafeRoutes() + " network_category: public\n")
require.NoError(t, res.Err)
assert.Equal(t, []string{"tun.network_category"}, res.RestartKeys)
}
// forFamilies runs f with two unsafe_routes CIDRs of each address family, the tun's own network stays IPv4
func forFamilies(t *testing.T, f func(t *testing.T, a, b string)) {
t.Run("ipv4", func(t *testing.T) { f(t, "10.251.1.0/24", "10.251.2.0/24") })
t.Run("ipv6", func(t *testing.T) { f(t, "fd00:251:1::/64", "fd00:251:2::/64") })
}
+70 -18
View File
@@ -2,7 +2,9 @@ package nebula
import (
"crypto/fips140"
"crypto/sha256"
"encoding/binary"
"encoding/hex"
"encoding/json"
"errors"
"fmt"
@@ -50,15 +52,30 @@ type CertState struct {
myVpnBroadcastAddrsTable *bart.Lite
}
// errCertNeedsRestart marks a new cert that is refused on reload because only a restart can apply it
var errCertNeedsRestart = errors.New("a restart is required to use the new cert")
func NewPKIFromConfig(l *slog.Logger, c *config.C) (*PKI, error) {
pki := &PKI{l: l}
err := pki.reload(c, true)
err := pki.reload(c, true, nil)
if err != nil {
return nil, err
}
c.RegisterReloadCallback(func(c *config.C) {
rErr := pki.reload(c, false)
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
// Compare with the running cipher, not the last config. A restart with one we can't run would fail, so say that
// now rather than ask for one, and going back to the running one after a bad reload needs no restart
if cipher := c.GetString("cipher", "aes"); !validCipher(cipher) {
l.Error("Unknown cipher, a restart with it would fail", "cipher", cipher)
r.Failed()
} else if cipher == "chachapoly" && fipsOnly() {
// newCipherSuite refuses it at startup
l.Error("ChaChaPoly is not allowed in FIPS 140-only mode, a restart with it would fail", "cipher", cipher)
r.Failed()
} else if cipher != pki.getCertState().cipher {
r.RestartRequiredIfChanged("cipher")
}
rErr := pki.reload(c, false, r)
if rErr != nil {
util.LogWithContextIfNeeded("Failed to reload PKI from config", rErr, l)
}
@@ -75,13 +92,29 @@ func (p *PKI) getCertState() *CertState {
return p.cs.Load()
}
func (p *PKI) reload(c *config.C, initial bool) error {
// r is nil on the initial load
func (p *PKI) reload(c *config.C, initial bool, r *config.Reload) error {
before := p.cs.Load()
err := p.reloadCerts(c, initial)
if err != nil {
if initial {
return err
}
err.Log(p.l)
if errors.Is(err, errCertNeedsRestart) {
r.RestartRequired()
} else {
r.Failed()
}
} else if !initial {
// A cert for a version we weren't running is accepted without comparing its networks to ours. The tun and the
// lighthouse keep the networks they started with
after := p.cs.Load()
if !slices.Equal(before.myVpnNetworks, after.myVpnNetworks) {
p.l.Warn("The new cert changes our networks, a restart is required to use them",
"old_networks", before.myVpnNetworks, "new_networks", after.myVpnNetworks)
r.RestartRequired()
}
}
err = p.reloadCAPool(c)
@@ -90,21 +123,27 @@ func (p *PKI) reload(c *config.C, initial bool) error {
return err
}
err.Log(p.l)
r.Failed()
}
return nil
}
// fipsOnly is fips140.Enforced, a variable so a test can run the FIPS 140-only reload check without GODEBUG
var fipsOnly = fips140.Enforced
// validCipher is true for the ciphers nebula can run. Each post-handshake CipherState in noiseutil hardcodes its own
// nonce endianness, so there's nothing to set up for either.
func validCipher(cipher string) bool {
return cipher == "aes" || cipher == "chachapoly"
}
func (p *PKI) reloadCerts(c *config.C, initial bool) *util.ContextualError {
var cipher string
var currentState *CertState
if initial {
cipher = c.GetString("cipher", "aes")
switch cipher {
case "aes", "chachapoly":
// Each post-handshake CipherState in noiseutil hardcodes its own
// nonce endianness now, so there's nothing to set up here.
default:
if !validCipher(cipher) {
return util.NewContextualError(
"unknown cipher",
m{"cipher": cipher},
@@ -132,7 +171,7 @@ func (p *PKI) reloadCerts(c *config.C, initial bool) *util.ContextualError {
return util.NewContextualError(
"Networks in new cert was different from old",
m{"new_networks": newState.v1Cert.Networks(), "old_networks": currentState.v1Cert.Networks(), "cert_version": cert.Version1},
nil,
errCertNeedsRestart,
)
}
@@ -140,7 +179,7 @@ func (p *PKI) reloadCerts(c *config.C, initial bool) *util.ContextualError {
return util.NewContextualError(
"Curve in new v1 cert was different from old",
m{"new_curve": newState.v1Cert.Curve(), "old_curve": currentState.v1Cert.Curve(), "cert_version": cert.Version1},
nil,
errCertNeedsRestart,
)
}
}
@@ -155,7 +194,7 @@ func (p *PKI) reloadCerts(c *config.C, initial bool) *util.ContextualError {
return util.NewContextualError(
"Networks in new cert was different from old",
m{"new_networks": newState.v2Cert.Networks(), "old_networks": currentState.v2Cert.Networks(), "cert_version": cert.Version2},
nil,
errCertNeedsRestart,
)
}
@@ -163,7 +202,7 @@ func (p *PKI) reloadCerts(c *config.C, initial bool) *util.ContextualError {
return util.NewContextualError(
"Curve in new cert was different from old",
m{"new_curve": newState.v2Cert.Curve(), "old_curve": currentState.v2Cert.Curve(), "cert_version": cert.Version2},
nil,
errCertNeedsRestart,
)
}
}
@@ -178,7 +217,7 @@ func (p *PKI) reloadCerts(c *config.C, initial bool) *util.ContextualError {
return util.NewContextualError(
"Removing a V2 cert is not permitted unless it has identical networks to the new V1 cert",
m{"new_v1_networks": newState.v1Cert.Networks(), "old_v2_networks": currentState.v2Cert.Networks()},
nil,
errCertNeedsRestart,
)
}
}
@@ -196,12 +235,14 @@ func (p *PKI) reloadCerts(c *config.C, initial bool) *util.ContextualError {
func (p *PKI) reloadCAPool(c *config.C) *util.ContextualError {
caPool, err := loadCAPoolFromConfig(p.l, c)
if caPool != nil {
p.caPool.Store(caPool)
p.l.Debug("Trusted CA fingerprints", "fingerprints", caPool.GetFingerprints())
}
if err != nil {
return util.NewContextualError("Failed to load ca from config", nil, err)
}
p.caPool.Store(caPool)
p.l.Debug("Trusted CA fingerprints", "fingerprints", caPool.GetFingerprints())
return nil
}
@@ -576,11 +617,22 @@ func loadCAPoolFromConfig(l *slog.Logger, c *config.C) (*cert.CAPool, error) {
bl := c.GetStringSlice("pki.blocklist", []string{})
if len(bl) > 0 {
for _, fp := range bl {
var bad []string
for i, raw := range bl {
// Fingerprints are compared as lowercase hex strings, a malformed entry would silently block nothing
fp := strings.ToLower(strings.TrimSpace(raw))
if b, err := hex.DecodeString(fp); err != nil || len(b) != sha256.Size {
bad = append(bad, fmt.Sprintf("entry %d %q", i+1, raw))
continue
}
caPool.BlocklistFingerprint(fp)
}
l.Info("Blocklisted certificates", "fingerprintCount", len(bl))
l.Info("Blocklisted certificates", "fingerprintCount", len(bl)-len(bad))
if len(bad) > 0 {
// The valid entries are still returned so a failed reload blocks them, startup refuses the error
return caPool, fmt.Errorf("pki.blocklist entries are not hex certificate fingerprints: %s", strings.Join(bad, ", "))
}
}
return caPool, nil
+168
View File
@@ -1,10 +1,17 @@
package nebula
import (
"crypto/fips140"
"fmt"
"net/netip"
"strings"
"testing"
"time"
"github.com/slackhq/nebula/cert"
"github.com/slackhq/nebula/cert_test"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/test"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
@@ -92,3 +99,164 @@ func TestNewCipherSuiteUnsupportedCurve(t *testing.T) {
assert.True(t, strings.HasPrefix(err.Error(), "unsupported curve:"), "got: %v", err)
assert.Nil(t, cs)
}
func TestPKIReloadRestartRequired(t *testing.T) {
ca, _, caKey, caPEM := cert_test.NewTestCaCert(cert.Version2, cert.Curve_CURVE25519, time.Time{}, time.Time{}, nil, nil, nil)
pkiConfig := func(cipher, network string) string {
_, _, keyPEM, certPEM := cert_test.NewTestCert(cert.Version2, cert.Curve_CURVE25519, ca, caKey, "host", time.Time{}, time.Time{}, []netip.Prefix{netip.MustParsePrefix(network)}, nil, nil)
return fmt.Sprintf("cipher: %s\npki:\n ca: %q\n cert: %q\n key: %q\n", cipher, caPEM, certPEM, keyPEM)
}
c := config.NewC(test.NewLogger())
require.NoError(t, c.LoadString(pkiConfig("aes", "10.1.0.2/24")))
p, err := NewPKIFromConfig(test.NewLogger(), c)
require.NoError(t, err)
// A cert that changes our networks is refused and needs a restart, it is not a failure
res := c.ReloadConfigString(pkiConfig("aes", "10.1.0.3/24"))
require.NoError(t, res.Err)
assert.True(t, res.RestartRequired)
assert.False(t, res.Failed)
assert.Equal(t, []netip.Prefix{netip.MustParsePrefix("10.1.0.2/24")}, p.getCertState().myVpnNetworks)
// The cipher only applies at startup
c2 := config.NewC(test.NewLogger())
require.NoError(t, c2.LoadString(pkiConfig("aes", "10.1.0.2/24")))
p, err = NewPKIFromConfig(test.NewLogger(), c2)
require.NoError(t, err)
res = c2.ReloadConfigString(pkiConfig("chachapoly", "10.1.0.2/24"))
require.NoError(t, res.Err)
assert.True(t, res.RestartRequired)
assert.Equal(t, "aes", p.getCertState().cipher)
// A cipher we can't run would fail the restart, so it fails the reload instead of asking for one
c4 := config.NewC(test.NewLogger())
require.NoError(t, c4.LoadString(pkiConfig("aes", "10.1.0.2/24")))
_, err = NewPKIFromConfig(test.NewLogger(), c4)
require.NoError(t, err)
for range 2 {
res = c4.ReloadConfigString(pkiConfig("nope", "10.1.0.2/24"))
require.NoError(t, res.Err)
assert.True(t, res.Failed, "a bad cipher that didn't change since the last reload is still bad")
assert.False(t, res.RestartRequired)
}
// Fixing it back to the running cipher needs nothing
res = c4.ReloadConfigString(pkiConfig("aes", "10.1.0.2/24"))
require.NoError(t, res.Err)
assert.False(t, res.Failed)
assert.False(t, res.RestartRequired)
// A cert that keeps our networks applies without a restart
c3 := config.NewC(test.NewLogger())
require.NoError(t, c3.LoadString(pkiConfig("aes", "10.1.0.2/24")))
_, err = NewPKIFromConfig(test.NewLogger(), c3)
require.NoError(t, err)
res = c3.ReloadConfigString(pkiConfig("aes", "10.1.0.2/24"))
require.NoError(t, res.Err)
assert.False(t, res.RestartRequired)
// A cert that can't be read is a failure, and a restart wouldn't help
res = c3.ReloadConfigString(strings.Replace(pkiConfig("aes", "10.1.0.2/24"), "cert: \"", "cert: \"garbage", 1))
require.NoError(t, res.Err)
assert.True(t, res.Failed)
assert.False(t, res.RestartRequired)
}
// A cert for a version we weren't running is accepted without comparing networks, so switching to one with different
// networks needs a restart. One with the same networks doesn't
func TestPKIReloadNewVersionWithOtherNetworks(t *testing.T) {
ca1, _, ca1Key, ca1PEM := cert_test.NewTestCaCert(cert.Version1, cert.Curve_CURVE25519, time.Time{}, time.Time{}, nil, nil, nil)
ca2, _, ca2Key, ca2PEM := cert_test.NewTestCaCert(cert.Version2, cert.Curve_CURVE25519, time.Time{}, time.Time{}, nil, nil, nil)
caPEM := string(ca1PEM) + string(ca2PEM)
pkiConfig := func(v cert.Version, network string) string {
ca, caKey := ca1, ca1Key
if v == cert.Version2 {
ca, caKey = ca2, ca2Key
}
_, _, keyPEM, certPEM := cert_test.NewTestCert(v, cert.Curve_CURVE25519, ca, caKey, "host", time.Time{}, time.Time{}, []netip.Prefix{netip.MustParsePrefix(network)}, nil, nil)
return fmt.Sprintf("pki:\n ca: %q\n cert: %q\n key: %q\n", caPEM, certPEM, keyPEM)
}
for _, tc := range []struct {
network string
restart bool
}{
{"10.9.0.7/24", true},
{"10.1.0.2/24", false},
} {
c := config.NewC(test.NewLogger())
require.NoError(t, c.LoadString(pkiConfig(cert.Version1, "10.1.0.2/24")))
p, err := NewPKIFromConfig(test.NewLogger(), c)
require.NoError(t, err)
res := c.ReloadConfigString(pkiConfig(cert.Version2, tc.network))
require.NoError(t, res.Err)
assert.False(t, res.Failed, tc.network)
assert.Equal(t, tc.restart, res.RestartRequired, tc.network)
assert.Equal(t, []netip.Prefix{netip.MustParsePrefix(tc.network)}, p.getCertState().myVpnNetworks, "accepted either way, as before")
}
}
// FIPS 140-only mode refuses chachapoly at startup, so reloading to it fails rather than asks for a restart
func TestPKIReloadChachapolyInFIPSOnlyFails(t *testing.T) {
fipsOnly = func() bool { return true }
t.Cleanup(func() { fipsOnly = fips140.Enforced })
ca, _, caKey, caPEM := cert_test.NewTestCaCert(cert.Version2, cert.Curve_P256, time.Time{}, time.Time{}, nil, nil, nil)
_, _, keyPEM, certPEM := cert_test.NewTestCert(cert.Version2, cert.Curve_P256, ca, caKey, "host", time.Time{}, time.Time{}, []netip.Prefix{netip.MustParsePrefix("10.1.0.2/24")}, nil, nil)
pkiConfig := func(cipher string) string {
return fmt.Sprintf("cipher: %s\npki:\n ca: %q\n cert: %q\n key: %q\n", cipher, caPEM, certPEM, keyPEM)
}
c := config.NewC(test.NewLogger())
require.NoError(t, c.LoadString(pkiConfig("aes")))
_, err := NewPKIFromConfig(test.NewLogger(), c)
require.NoError(t, err)
res := c.ReloadConfigString(pkiConfig("chachapoly"))
require.NoError(t, res.Err)
assert.True(t, res.Failed)
assert.False(t, res.RestartRequired)
}
func TestLoadCAPoolBlocklist(t *testing.T) {
ca, _, caKey, caPEM := cert_test.NewTestCaCert(cert.Version2, cert.Curve_CURVE25519, time.Time{}, time.Time{}, nil, nil, nil)
crt, _, _, _ := cert_test.NewTestCert(cert.Version2, cert.Curve_CURVE25519, ca, caKey, "host", time.Time{}, time.Time{}, []netip.Prefix{netip.MustParsePrefix("10.1.0.2/24")}, nil, nil)
fp, err := crt.Fingerprint()
require.NoError(t, err)
load := func(entry string) (*cert.CAPool, error) {
c := config.NewC(test.NewLogger())
require.NoError(t, c.LoadString(fmt.Sprintf("pki:\n ca: %q\n blocklist:\n - %q\n", caPEM, entry)))
return loadCAPoolFromConfig(test.NewLogger(), c)
}
for _, entry := range []string{fp, strings.ToUpper(fp), " " + fp + "\t"} {
pool, err := load(entry)
require.NoError(t, err, entry)
assert.True(t, pool.IsBlocklisted(fp), entry)
}
for _, entry := range []string{fp[1:], fp + "00", "sha256:" + fp, "zz" + fp[2:], ""} {
_, err := load(entry)
require.Error(t, err, entry)
}
// A bad entry on reload fails it, the valid entries are still blocked
_, _, keyPEM, certPEM := cert_test.NewTestCert(cert.Version2, cert.Curve_CURVE25519, ca, caKey, "me", time.Time{}, time.Time{}, []netip.Prefix{netip.MustParsePrefix("10.1.0.1/24")}, nil, nil)
pkiConfig := func(entry string) string {
return fmt.Sprintf("pki:\n ca: %q\n cert: %q\n key: %q\n blocklist:\n - %q\n", caPEM, certPEM, keyPEM, entry)
}
c := config.NewC(test.NewLogger())
require.NoError(t, c.LoadString(pkiConfig(fp)))
p, err := NewPKIFromConfig(test.NewLogger(), c)
require.NoError(t, err)
other, _, _, _ := cert_test.NewTestCert(cert.Version2, cert.Curve_CURVE25519, ca, caKey, "other", time.Time{}, time.Time{}, []netip.Prefix{netip.MustParsePrefix("10.1.0.3/24")}, nil, nil)
otherFp, err := other.Fingerprint()
require.NoError(t, err)
res := c.ReloadConfigString(pkiConfig(fp) + fmt.Sprintf(" - %q\n - %q\n", otherFp, fp[1:]))
assert.True(t, res.Failed)
assert.True(t, p.GetCAPool().IsBlocklisted(fp))
assert.True(t, p.GetCAPool().IsBlocklisted(otherFp))
}
+2 -1
View File
@@ -67,7 +67,8 @@ func NewPunchyFromConfig(l *slog.Logger, c *config.C, punchConn udp.Conn) *Punch
}
p.reload(c, true)
c.RegisterReloadCallback(func(c *config.C) {
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
r.RestartRequiredIfChanged("stats.lighthouse_metrics")
p.reload(c, false)
})
+16 -6
View File
@@ -76,7 +76,7 @@ punchy:
punch: true
delay: 10m
respond: true
`))
`).Err)
p.reload(c, false)
assert.True(t, p.punch.Load())
assert.Equal(t, newDelay, time.Duration(p.delay.Load()))
@@ -170,7 +170,7 @@ func TestPunchy_LogFormat_ReloadPunch(t *testing.T) {
NewPunchyFromConfig(l, c, nil)
hook.entries = nil
require.NoError(t, c.ReloadConfigString(`punchy: {punch: true}`))
require.NoError(t, c.ReloadConfigString(`punchy: {punch: true}`).Err)
entry := findEntry(t, hook.entries, "punchy.punch changed")
assert.Equal(t, slog.LevelInfo, entry.Level)
@@ -184,7 +184,7 @@ func TestPunchy_LogFormat_ReloadRespond(t *testing.T) {
NewPunchyFromConfig(l, c, nil)
hook.entries = nil
require.NoError(t, c.ReloadConfigString(`punchy: {respond: true}`))
require.NoError(t, c.ReloadConfigString(`punchy: {respond: true}`).Err)
entry := findEntry(t, hook.entries, "punchy.respond changed")
assert.Equal(t, slog.LevelInfo, entry.Level)
@@ -198,7 +198,7 @@ func TestPunchy_LogFormat_ReloadDelay(t *testing.T) {
NewPunchyFromConfig(l, c, nil)
hook.entries = nil
require.NoError(t, c.ReloadConfigString(`punchy: {delay: 10s}`))
require.NoError(t, c.ReloadConfigString(`punchy: {delay: 10s}`).Err)
entry := findEntry(t, hook.entries, "punchy.delay changed")
assert.Equal(t, slog.LevelInfo, entry.Level)
@@ -212,7 +212,7 @@ func TestPunchy_LogFormat_ReloadTargetAllRemotes(t *testing.T) {
NewPunchyFromConfig(l, c, nil)
hook.entries = nil
require.NoError(t, c.ReloadConfigString(`punchy: {target_all_remotes: true}`))
require.NoError(t, c.ReloadConfigString(`punchy: {target_all_remotes: true}`).Err)
entry := findEntry(t, hook.entries, "punchy.target_all_remotes changed")
assert.Equal(t, slog.LevelInfo, entry.Level)
@@ -226,9 +226,19 @@ func TestPunchy_LogFormat_ReloadRespondDelay(t *testing.T) {
NewPunchyFromConfig(l, c, nil)
hook.entries = nil
require.NoError(t, c.ReloadConfigString(`punchy: {respond_delay: 15s}`))
require.NoError(t, c.ReloadConfigString(`punchy: {respond_delay: 15s}`).Err)
entry := findEntry(t, hook.entries, "punchy.respond_delay changed")
assert.Equal(t, slog.LevelInfo, entry.Level)
assert.Equal(t, map[string]any{"respond_delay": 15 * time.Second}, entry.Attrs)
}
func TestPunchyReloadRestartRequired(t *testing.T) {
c := config.NewC(test.NewLogger())
require.NoError(t, c.LoadString("stats:\n lighthouse_metrics: false\n"))
NewPunchyFromConfig(test.NewLogger(), c, nil)
res := c.ReloadConfigString("stats:\n lighthouse_metrics: true\n")
require.NoError(t, res.Err)
assert.True(t, res.RestartRequired)
}
+91
View File
@@ -0,0 +1,91 @@
package ratelimit
import (
"fmt"
"math"
"sync/atomic"
"time"
)
// Limiter is a token bucket: `burst` calls are allowed at once, then one more per `every`.
// It counts the calls it denies and hands that count to the next allowed call.
//
// The budget belongs to the Limiter instance.
// Every Allow that returns true spends from it, whatever the caller then does.
//
// Time follows Go's monotonic clock.
//
// Create one Limiter for each thing you want to limit,
// sharing one between unrelated events lets the noisy one starve the quiet one.
//
// Check Allow before doing the work being limited, so a denied call stays cheap:
//
// if denied, ok := probeLimiter.Allow(); ok {
// sendProbe()
// probesDenied.Inc(int64(denied))
// }
//
// For log lines use logging.RateLimited, which attaches the denied count to the logger as "suppressed".
type Limiter struct {
every int64
slack int64
tat atomic.Int64
denied atomic.Uint64
}
// maxSpan bounds `burst` * `every` to half of int64, leaving the other half for process uptime so the bucket
// math can't overflow.
const maxSpan = math.MaxInt64 / 2
// New returns a Limiter allowing `burst` calls at once and one more per `every`.
// A `burst` under 1 is 1, and an `every` of 0 or less never limits.
// It returns an error if `burst` * `every` is over about 146 years.
func New(burst int, every time.Duration) (*Limiter, error) {
l := &Limiter{}
if every > 0 {
burst = max(burst, 1)
if int64(burst) > maxSpan/int64(every) {
return nil, fmt.Errorf("ratelimit: burst %d * every %v is too long", burst, every)
}
l.every = int64(every)
l.slack = int64(burst-1) * l.every
}
return l, nil
}
var monoStart = time.Now()
func monoNow() int64 { return int64(time.Since(monoStart)) }
// Allow reports whether this call is allowed.
// If a token is available it spends it and returns the calls denied since the last allowed one, and true.
// Otherwise it returns 0 and false, and counts the denial.
// A nil Limiter allows everything.
func (l *Limiter) Allow() (uint64, bool) {
if l == nil {
return 0, true
}
return l.allow(monoNow())
}
// allow is Allow at a given monotonic time.
func (l *Limiter) allow(now int64) (uint64, bool) {
// Concurrent Allow calls read the clock before racing for the bucket, so one can see a time another already spent
// past and be denied. An unlimited Limiter must never deny, so it skips the bucket.
if l.every == 0 {
return 0, true
}
for {
tat := l.tat.Load()
next := max(tat, now)
if next-now > l.slack {
l.denied.Add(1)
return 0, false
}
if l.tat.CompareAndSwap(tat, next+l.every) {
return l.denied.Swap(0), true
}
}
}
+264
View File
@@ -0,0 +1,264 @@
package ratelimit
import (
"math"
"sync"
"sync/atomic"
"testing"
"time"
)
func mustLimiter(t testing.TB, burst int, every time.Duration) *Limiter {
t.Helper()
l, err := New(burst, every)
if err != nil {
t.Fatal(err)
}
return l
}
func TestLimiterBurstAndRefill(t *testing.T) {
l, now := mustLimiter(t, 3, time.Second), int64(0)
for i := 0; i < 3; i++ {
if n, ok := l.allow(now); !ok || n != 0 {
t.Fatalf("burst line %d: ok=%v denied=%d", i, ok, n)
}
}
for i := 0; i < 5; i++ {
if _, ok := l.allow(now); ok {
t.Fatalf("line %d past the burst logged", i)
}
}
now += int64(time.Second)
if n, ok := l.allow(now); !ok || n != 5 {
t.Fatalf("after one refill: ok=%v denied=%d, want true 5", ok, n)
}
if _, ok := l.allow(now); ok {
t.Fatal("one refill earned more than one line")
}
// A long quiet spell refills to burst, not past it, and the count reported above does not carry over.
now += int64(time.Hour)
for i, want := range []uint64{1, 0, 0} {
if n, ok := l.allow(now); !ok || n != want {
t.Fatalf("refilled line %d: ok=%v denied=%d, want true %d", i, ok, n, want)
}
}
if _, ok := l.allow(now); ok {
t.Fatal("bucket refilled past burst")
}
}
func TestLimiterBounds(t *testing.T) {
for _, every := range []time.Duration{0, -time.Second} {
l := mustLimiter(t, 1, every)
for i := 0; i < 100; i++ {
if _, ok := l.allow(0); !ok {
t.Fatalf("every=%v limited line %d", every, i)
}
}
}
l := mustLimiter(t, 0, time.Second)
if _, ok := l.allow(0); !ok {
t.Fatal("burst 0 should allow one line")
}
if _, ok := l.allow(0); ok {
t.Fatal("burst 0 should allow only one line")
}
}
func TestLimiterOverflow(t *testing.T) {
for _, c := range []struct {
burst int
every time.Duration
}{
{math.MaxInt, time.Second},
{1, math.MaxInt64},
{2, maxSpan/2 + 1},
} {
if _, err := New(c.burst, c.every); err == nil {
t.Fatalf("burst %d every %d should not fit", c.burst, c.every)
}
}
l := mustLimiter(t, 2, maxSpan/2)
for i := 0; i < 2; i++ {
if _, ok := l.allow(0); !ok {
t.Fatalf("the largest span should allow its burst, denied call %d", i)
}
}
if _, ok := l.allow(0); ok {
t.Fatal("the largest span allowed past its burst")
}
}
// Concurrent callers read the clock at slightly different times, which must not limit a Limiter that never limits.
func TestLimiterUnlimitedConcurrent(t *testing.T) {
l := mustLimiter(t, 1, 0)
var denied atomic.Int64
var wg sync.WaitGroup
for i := 0; i < 16; i++ {
wg.Add(1)
go func() {
defer wg.Done()
for j := 0; j < 10000; j++ {
if _, ok := l.Allow(); !ok {
denied.Add(1)
}
}
}()
}
wg.Wait()
if got := denied.Load(); got != 0 {
t.Fatalf("denied %d lines", got)
}
}
func TestLimiterNil(t *testing.T) {
var l *Limiter
for i := 0; i < 10; i++ {
if _, ok := l.Allow(); !ok {
t.Fatal("nil limiter limited")
}
}
}
func TestLimiterConcurrent(t *testing.T) {
l := mustLimiter(t, 10, time.Hour)
var allowed atomic.Int64
var reported atomic.Uint64
var wg sync.WaitGroup
for i := 0; i < 16; i++ {
wg.Add(1)
go func() {
defer wg.Done()
for j := 0; j < 1000; j++ {
if n, ok := l.allow(0); ok {
allowed.Add(1)
reported.Add(n)
}
}
}()
}
wg.Wait()
if got := allowed.Load(); got != 10 {
t.Fatalf("allowed %d lines, want the burst of 10", got)
}
// An allowed call that lands after some denials reports them, so the rest are still pending.
if got := reported.Load() + l.denied.Load(); got != 16*1000-10 {
t.Fatalf("denied %d, want %d", got, 16*1000-10)
}
}
func BenchmarkAllowAllowed(b *testing.B) {
l := mustLimiter(b, 1, time.Nanosecond)
b.ReportAllocs()
for i := 0; i < b.N; i++ {
l.tat.Store(0)
l.Allow()
}
}
func BenchmarkAllowDenied(b *testing.B) {
l := mustLimiter(b, 1, time.Hour)
l.Allow()
b.ReportAllocs()
for i := 0; i < b.N; i++ {
l.Allow()
}
}
func BenchmarkAllowDeniedParallel(b *testing.B) {
l := mustLimiter(b, 1, time.Hour)
l.Allow()
b.ReportAllocs()
b.RunParallel(func(pb *testing.PB) {
for pb.Next() {
l.Allow()
}
})
}
// FuzzLimiter checks the bucket against its exact definition over arbitrary call timings.
// A call is denied if and only if some allowed call a would leave too many calls for the time since a:
// (allowed calls since a, plus this one) * every > burst * every + (now - a).
// Every denied call must also be reported exactly once, by a later allowed call or still pending.
func FuzzLimiter(f *testing.F) {
f.Add(uint8(3), uint16(1000), []byte{0, 0, 0, 0, 16, 0, 0, 200, 1, 1, 1, 15, 17, 0})
f.Add(uint8(1), uint16(1), []byte{0, 16, 15, 16, 0, 255, 0, 0})
f.Fuzz(func(t *testing.T, burst uint8, everyUs uint16, steps []byte) {
if len(steps) > 512 {
steps = steps[:512]
}
b := int64(burst%16) + 1
every := int64(everyUs%1000+1) * int64(time.Microsecond)
l := mustLimiter(t, int(b), time.Duration(every))
var now int64
var allowedAt []int64
var denied, reported uint64
for _, s := range steps {
// Each step advances by a sixteenth of every, so timings land on and around the refill.
now += int64(s) * every / 16
over := false
for i, a := range allowedAt {
if int64(len(allowedAt)-i+1)*every > b*every+now-a {
over = true
break
}
}
n, ok := l.allow(now)
if ok == over {
t.Fatalf("at %d with %d allowed before: ok=%v, the definition says %v", now, len(allowedAt), ok, !over)
}
if !ok {
denied++
continue
}
reported += n
allowedAt = append(allowedAt, now)
}
if got := reported + l.denied.Load(); got != denied {
t.Fatalf("reported %d denied calls, made %d", got, denied)
}
})
}
// Concurrent callers on the real clock must account for every denied call exactly once.
func TestLimiterDeniedConservedConcurrent(t *testing.T) {
l := mustLimiter(t, 4, time.Microsecond)
var denied, reported atomic.Uint64
var wg sync.WaitGroup
for i := 0; i < 16; i++ {
wg.Add(1)
go func() {
defer wg.Done()
for j := 0; j < 20000; j++ {
if n, ok := l.Allow(); ok {
reported.Add(n)
} else {
denied.Add(1)
}
}
}()
}
wg.Wait()
if denied.Load() == 0 {
t.Fatal("nothing was denied, the test proves nothing")
}
if got := reported.Load() + l.denied.Load(); got != denied.Load() {
t.Fatalf("reported %d denied calls, made %d", got, denied.Load())
}
}
+2 -1
View File
@@ -28,10 +28,11 @@ func NewRelayManager(ctx context.Context, l *slog.Logger, hostmap *HostMap, c *c
hostmap: hostmap,
}
rm.reload(c, true)
c.RegisterReloadCallback(func(c *config.C) {
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
err := rm.reload(c, false)
if err != nil {
rm.l.Error("Failed to reload relay_manager", "error", err)
r.Failed()
}
})
return rm
+26 -9
View File
@@ -1,8 +1,8 @@
package service
import (
"io"
"net"
"sync"
)
type tcpListener struct {
@@ -10,23 +10,40 @@ type tcpListener struct {
s *Service
addr *net.TCPAddr
accept chan net.Conn
// closed is closed instead of accept so a handler blocked handing off a
// connection can't send on a closed channel.
closed chan struct{}
closeOnce sync.Once
}
func (l *tcpListener) Accept() (net.Conn, error) {
conn, ok := <-l.accept
if !ok {
return nil, io.EOF
// select picks randomly among ready cases, so without this a handler still
// waiting to hand off could win over closed after Close has returned.
select {
case <-l.closed:
return nil, net.ErrClosed
default:
}
select {
case conn := <-l.accept:
return conn, nil
case <-l.closed:
return nil, net.ErrClosed
}
}
func (l *tcpListener) Close() error {
l.closeOnce.Do(func() {
// Close before leaving the map so a concurrent closeListeners that
// no longer sees this listener knows Accept is already released.
close(l.closed)
l.s.mu.Lock()
defer l.s.mu.Unlock()
delete(l.s.mu.listeners, uint16(l.addr.Port))
close(l.accept)
delete(l.s.mu.listeners, l.port)
l.s.mu.Unlock()
})
return nil
}
+79 -10
View File
@@ -32,13 +32,18 @@ import (
const nicID = 1
type Service struct {
eg *errgroup.Group
control *nebula.Control
ipstack *stack.Stack
// done is closed once the forwarding goroutines have exited and the
// netstack is torn down, at which point err holds the first fatal error.
done chan struct{}
err error
mu struct {
sync.Mutex
closed bool
listeners map[uint16]*tcpListener
}
}
@@ -65,7 +70,6 @@ func New(control *nebula.Control) (_ *Service, reterr error) {
ctx := control.Context()
eg, ctx := errgroup.WithContext(ctx)
s := Service{
eg: eg,
control: control,
}
s.mu.listeners = map[uint16]*tcpListener{}
@@ -74,6 +78,12 @@ func New(control *nebula.Control) (_ *Service, reterr error) {
NetworkProtocols: []stack.NetworkProtocolFactory{ipv4.NewProtocol, ipv6.NewProtocol},
TransportProtocols: []stack.TransportProtocolFactory{tcp.NewProtocol, udp.NewProtocol, icmp.NewProtocol4, icmp.NewProtocol6},
})
// The teardown goroutine only owns the stack once New succeeds
defer func() {
if reterr != nil {
s.ipstack.Destroy()
}
}()
sackEnabledOpt := tcpip.TCPSACKEnabled(true) // TCP SACK is disabled by default
tcpipErr := s.ipstack.SetTransportProtocolOption(tcp.ProtocolNumber, &sackEnabledOpt)
if tcpipErr != nil {
@@ -123,6 +133,12 @@ func New(control *nebula.Control) (_ *Service, reterr error) {
// this will read exactly one packet
n, err := reader.Read(buf)
if err != nil {
// An error after shutdown began is noise from the device pipes
// closing, anything else is fatal and must stop nebula too
if ctx.Err() != nil {
return nil
}
s.control.Stop()
return err
}
packetBuf := stack.NewPacketBuffer(stack.PacketBufferOptions{
@@ -130,8 +146,8 @@ func New(control *nebula.Control) (_ *Service, reterr error) {
})
linkEP.InjectInbound(header.IPv4ProtocolNumber, packetBuf)
if err := ctx.Err(); err != nil {
return err
if ctx.Err() != nil {
return nil
}
}
})
@@ -139,13 +155,18 @@ func New(control *nebula.Control) (_ *Service, reterr error) {
for {
packet := linkEP.ReadContext(ctx)
if packet == nil {
if err := ctx.Err(); err != nil {
return err
if ctx.Err() != nil {
return nil
}
continue
}
bufView := packet.ToView()
if _, err := bufView.WriteTo(writer); err != nil {
// Same shutdown noise handling as the inbound reader
if ctx.Err() != nil {
return nil
}
s.control.Stop()
return err
}
bufView.Release()
@@ -158,6 +179,22 @@ func New(control *nebula.Control) (_ *Service, reterr error) {
return control.Wait()
})
// Tear down here rather than in Close so it also happens when nebula stops
// on its own. The forwarding goroutines must exit before the stack they feed
// is destroyed.
s.done = make(chan struct{})
go func() {
s.err = eg.Wait()
s.closeListeners()
// Destroy aborts the endpoints that exist when it starts, then waits on
// every endpoint, so a Dial registering in between would hold Wait until
// its connect timed out. Without the NIC new dials fail route lookup.
// RemoveNIC also closes linkEP.
_ = s.ipstack.RemoveNIC(nicID)
s.ipstack.Destroy()
close(s.done)
}()
return &s, nil
}
@@ -231,11 +268,15 @@ func (s *Service) Listen(network, address string) (net.Listener, error) {
s: s,
addr: addr,
accept: make(chan net.Conn),
closed: make(chan struct{}),
}
s.mu.Lock()
defer s.mu.Unlock()
if s.mu.closed {
return nil, net.ErrClosed
}
if _, ok := s.mu.listeners[port]; ok {
return nil, fmt.Errorf("already listening on port %d", port)
}
@@ -244,22 +285,44 @@ func (s *Service) Listen(network, address string) (net.Listener, error) {
return l, nil
}
// Wait blocks until the service has fully stopped, either via Close or a fatal
// error, and returns the first fatal error if there was one.
func (s *Service) Wait() error {
return s.eg.Wait()
<-s.done
return s.err
}
// Close closes all listeners and stops nebula. Use Wait to block until
// shutdown has completed.
func (s *Service) Close() error {
s.closeListeners()
s.control.Stop()
return nil
}
// closeListeners unblocks pending Accept calls and rejects new listeners.
func (s *Service) closeListeners() {
// Every caller closes every listener rather than returning early once
// closed is set, so none returns while another is still closing them.
s.mu.Lock()
s.mu.closed = true
listeners := make([]*tcpListener, 0, len(s.mu.listeners))
for _, l := range s.mu.listeners {
listeners = append(listeners, l)
}
s.mu.Unlock()
for _, l := range listeners {
_ = l.Close()
}
}
func (s *Service) tcpHandler(r *tcp.ForwarderRequest) {
endpointID := r.ID()
s.mu.Lock()
defer s.mu.Unlock()
l, ok := s.mu.listeners[endpointID.LocalPort]
s.mu.Unlock()
if !ok {
r.Complete(true)
return
@@ -275,6 +338,12 @@ func (s *Service) tcpHandler(r *tcp.ForwarderRequest) {
r.Complete(false)
ep.SocketOptions().SetKeepAlive(true)
// The handoff blocks until Accept is called, so it must not hold s.mu or a
// listener nobody accepts from stalls every other port.
conn := gonet.NewTCPConn(&wq, ep)
l.accept <- conn
select {
case l.accept <- conn:
case <-l.closed:
_ = conn.Close()
}
}
+298
View File
@@ -188,3 +188,301 @@ func TestService(t *testing.T) {
t.Fatal(err)
}
}
// A listener that is never accepted from must not stall connections to other
// ports, and closing it must not block.
func TestServiceUnacceptedListener(t *testing.T) {
lighthousePort := ephemeralUDPPort(t)
ca, _, caKey, _ := cert_test.NewTestCaCert(cert.Version2, cert.Curve_CURVE25519, time.Now(), time.Now().Add(10*time.Minute), nil, nil, []string{})
a := newSimpleService(ca, caKey, "a", netip.MustParseAddr("10.0.0.1"), m{
"static_host_map": m{},
"lighthouse": m{
"am_lighthouse": true,
},
"listen": m{
"host": "0.0.0.0",
"port": lighthousePort,
},
})
t.Cleanup(func() { _ = a.Close() })
b := newSimpleService(ca, caKey, "b", netip.MustParseAddr("10.0.0.2"), m{
"static_host_map": m{
"10.0.0.1": []string{fmt.Sprintf("127.0.0.1:%d", lighthousePort)},
},
"lighthouse": m{
"hosts": []string{"10.0.0.1"},
"interval": 1,
},
})
t.Cleanup(func() { _ = b.Close() })
idle, err := a.Listen("tcp", ":1234")
if err != nil {
t.Fatal(err)
}
active, err := a.Listen("tcp", ":1235")
if err != nil {
t.Fatal(err)
}
ctx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
defer cancel()
idleConn, err := b.DialContext(ctx, "tcp", "10.0.0.1:1234")
if err != nil {
t.Fatal(err)
}
defer idleConn.Close()
accepted := make(chan error, 1)
go func() {
conn, err := active.Accept()
if err == nil {
err = conn.Close()
}
accepted <- err
}()
ctx, cancel = context.WithTimeout(context.Background(), 5*time.Second)
defer cancel()
activeConn, err := b.DialContext(ctx, "tcp", "10.0.0.1:1235")
if err != nil {
t.Fatalf("dial to accepted port stalled behind idle listener: %v", err)
}
defer activeConn.Close()
select {
case err := <-accepted:
if err != nil {
t.Fatal(err)
}
case <-time.After(5 * time.Second):
t.Fatal("accept stalled behind idle listener")
}
closed := make(chan error, 1)
go func() { closed <- idle.Close() }()
select {
case err := <-closed:
if err != nil {
t.Fatal(err)
}
case <-time.After(5 * time.Second):
t.Fatal("closing idle listener blocked")
}
if _, err := idle.Accept(); !errors.Is(err, net.ErrClosed) {
t.Fatalf("expected net.ErrClosed from closed listener, got %v", err)
}
if err := idle.Close(); err != nil {
t.Fatalf("second close: %v", err)
}
}
// Accept must report net.ErrClosed after Close even while a handler is still
// trying to hand off a connection.
func TestListenerAcceptAfterClose(t *testing.T) {
s := &Service{}
s.mu.listeners = map[uint16]*tcpListener{}
l := &tcpListener{
port: 1234,
s: s,
accept: make(chan net.Conn),
closed: make(chan struct{}),
}
s.mu.listeners[l.port] = l
if err := l.Close(); err != nil {
t.Fatal(err)
}
// Keep a send pending so every Accept has a connection ready alongside closed.
stop := make(chan struct{})
done := make(chan struct{})
go func() {
defer close(done)
for {
server, client := net.Pipe()
_ = client.Close()
select {
case l.accept <- server:
case <-stop:
_ = server.Close()
return
}
}
}()
defer func() {
close(stop)
<-done
}()
for range 100 {
time.Sleep(time.Millisecond)
conn, err := l.Accept()
if conn != nil {
_ = conn.Close()
}
if !errors.Is(err, net.ErrClosed) {
t.Fatalf("Accept after Close: got err %v, want net.ErrClosed", err)
}
}
}
// Close must release blocked Accept calls and let Wait report a clean shutdown.
func TestServiceClose(t *testing.T) {
ca, _, caKey, _ := cert_test.NewTestCaCert(cert.Version2, cert.Curve_CURVE25519, time.Now(), time.Now().Add(10*time.Minute), nil, nil, []string{})
a := newSimpleService(ca, caKey, "a", netip.MustParseAddr("10.0.0.1"), m{
"static_host_map": m{},
"listen": m{
"host": "127.0.0.1",
"port": 0,
},
})
ln, err := a.Listen("tcp", ":1234")
if err != nil {
t.Fatal(err)
}
accepted := make(chan error, 1)
go func() {
_, err := ln.Accept()
accepted <- err
}()
if err := a.Close(); err != nil {
t.Fatal(err)
}
select {
case err := <-accepted:
if !errors.Is(err, net.ErrClosed) {
t.Fatalf("expected net.ErrClosed from Accept, got %v", err)
}
case <-time.After(5 * time.Second):
t.Fatal("Accept still blocked after Close")
}
waited := make(chan error, 1)
go func() { waited <- a.Wait() }()
select {
case err := <-waited:
if err != nil {
t.Fatalf("expected clean shutdown from Wait, got %v", err)
}
case <-time.After(5 * time.Second):
t.Fatal("Wait still blocked after Close")
}
if a.ipstack.HasNIC(nicID) {
t.Fatal("netstack not torn down after Wait")
}
if err := a.Close(); err != nil {
t.Fatalf("second close: %v", err)
}
if _, err := a.Listen("tcp", ":1235"); !errors.Is(err, net.ErrClosed) {
t.Fatalf("expected net.ErrClosed from Listen after Close, got %v", err)
}
}
// If nebula stops without Close being called, the service must still release
// Accept and tear down the netstack.
func TestServiceNebulaStopped(t *testing.T) {
ca, _, caKey, _ := cert_test.NewTestCaCert(cert.Version2, cert.Curve_CURVE25519, time.Now(), time.Now().Add(10*time.Minute), nil, nil, []string{})
a := newSimpleService(ca, caKey, "a", netip.MustParseAddr("10.0.0.1"), m{
"static_host_map": m{},
"listen": m{
"host": "127.0.0.1",
"port": 0,
},
})
ln, err := a.Listen("tcp", ":1234")
if err != nil {
t.Fatal(err)
}
accepted := make(chan error, 1)
go func() {
_, err := ln.Accept()
accepted <- err
}()
a.control.Stop()
waited := make(chan error, 1)
go func() { waited <- a.Wait() }()
select {
case err := <-waited:
if err != nil {
t.Fatalf("expected clean shutdown from Wait, got %v", err)
}
case <-time.After(5 * time.Second):
t.Fatal("Wait still blocked after nebula stopped")
}
select {
case err := <-accepted:
if !errors.Is(err, net.ErrClosed) {
t.Fatalf("expected net.ErrClosed from Accept, got %v", err)
}
case <-time.After(5 * time.Second):
t.Fatal("Accept still blocked after nebula stopped")
}
if a.ipstack.HasNIC(nicID) {
t.Fatal("netstack not torn down after Wait")
}
}
// Dials racing Close must not hold Wait until their connect times out.
func TestServiceDialDuringClose(t *testing.T) {
ca, _, caKey, _ := cert_test.NewTestCaCert(cert.Version2, cert.Curve_CURVE25519, time.Now(), time.Now().Add(10*time.Minute), nil, nil, []string{})
a := newSimpleService(ca, caKey, "a", netip.MustParseAddr("10.0.0.1"), m{
"static_host_map": m{},
"listen": m{
"host": "127.0.0.1",
"port": 0,
},
})
stop := make(chan struct{})
var eg errgroup.Group
for range 8 {
eg.Go(func() error {
for {
select {
case <-stop:
return nil
default:
}
// 10.0.0.2 has no tunnel, so each dial sits in SYN-SENT until aborted
if c, err := a.Dial("tcp", "10.0.0.2:80"); err == nil {
_ = c.Close()
}
}
})
}
time.Sleep(50 * time.Millisecond)
if err := a.Close(); err != nil {
t.Fatal(err)
}
waited := make(chan error, 1)
go func() { waited <- a.Wait() }()
select {
case err := <-waited:
if err != nil {
t.Fatalf("expected clean shutdown from Wait, got %v", err)
}
case <-time.After(5 * time.Second):
t.Fatal("Wait blocked behind a dial that raced Close")
}
close(stop)
dialersDone := make(chan error, 1)
go func() { dialersDone <- eg.Wait() }()
select {
case <-dialersDone:
case <-time.After(5 * time.Second):
t.Fatal("Dial still blocked after Wait")
}
}
+56 -922
View File
File diff suppressed because it is too large. Load diff
+57
View File
@@ -0,0 +1,57 @@
package nebula
import (
"bytes"
"fmt"
"path/filepath"
"strings"
"testing"
"github.com/slackhq/nebula/commands"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/test"
)
// TestSSHCommandSurface is a characterization test for the ssh command surface.
// It pins the registered command set, their descriptions, help text and flag definitions
// so that moving the registry to a transport neutral package is provably a relocation
// rather than a rewrite. Regenerate with `go test . -run TestSSHCommandSurface -update`.
func TestSSHCommandSurface(t *testing.T) {
got := renderCommandSurface(registeredCommands(t).All())
test.Golden(t, filepath.Join("testdata", "ssh_commands.golden"), got)
}
// registeredCommands returns a registry holding every command nebula serves. The
// Interface is never dereferenced: attachCommands only closes over it.
func registeredCommands(t *testing.T) *commands.Registry {
t.Helper()
l := test.NewLogger()
reg := commands.New()
attachCommands(l, config.NewC(l), reg, &Interface{})
return reg
}
func renderCommandSurface(cmds []*commands.Command) string {
lines := make([]string, 0, len(cmds))
for _, cmd := range cmds {
b := &bytes.Buffer{}
fmt.Fprintf(b, "%s\n", cmd.Name)
fmt.Fprintf(b, " short: %s\n", cmd.ShortDescription)
fmt.Fprintf(b, " help: %s\n", cmd.Help)
if cmd.Flags != nil {
fl, _ := cmd.Flags()
if fl != nil {
fb := &bytes.Buffer{}
fl.SetOutput(fb)
fl.PrintDefaults()
fmt.Fprintf(b, " flags:\n%s", fb.String())
}
}
lines = append(lines, b.String())
}
// In the order All gives, so the golden pins display order too.
return strings.Join(lines, "")
}
+94
View File
@@ -0,0 +1,94 @@
package nebula
import (
"errors"
"io"
"os"
"path/filepath"
"testing"
"github.com/slackhq/nebula/commands"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/test"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
type lineWriter struct {
lines []string
err error
}
func (w *lineWriter) WriteLine(s string) error {
w.lines = append(w.lines, s)
return w.err
}
func (w *lineWriter) Write(s string) error { return w.WriteLine(s) }
func (w *lineWriter) WriteBytes(b []byte) error { return w.WriteLine(string(b)) }
func (w *lineWriter) GetWriter() io.Writer { return io.Discard }
func TestReloadCommandReportsTheResult(t *testing.T) {
dir := t.TempDir()
write := func(body string) {
require.NoError(t, os.WriteFile(filepath.Join(dir, "config.yml"), []byte(body), 0o600))
}
write("a: 1\nb: 1")
l := test.NewLogger()
c := config.NewC(l)
require.NoError(t, c.Load(dir))
fail := false
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
r.RestartRequiredIfChanged("b")
if fail {
r.Failed()
}
})
// reload returns what the client was told. A reload that failed is a commands.Failure
// to the surface, exit 1 over ssh, with the result line as its message; one that
// merely needs a restart is not.
reload := func(failed bool) []string {
w := &lineWriter{}
err := reloadCommand(l, c, w)
if failed {
var f *commands.Failure
require.ErrorAs(t, err, &f)
assert.Equal(t, w.lines[len(w.lines)-1], f.Msg)
} else {
require.NoError(t, err)
}
return w.lines
}
write("a: 2\nb: 1")
assert.Equal(t, []string{"Reloading config", "Config reload complete"}, reload(false))
fail = true
write("a: 3\nb: 1")
assert.Equal(t, []string{"Reloading config", "Config reload completed with errors, review the logs"}, reload(true))
fail = false
write("a: 3\nb: 2")
assert.Equal(t, []string{"Reloading config", "Config reload was completed but requires a restart to fully apply, keys [b]"}, reload(false))
write("a: [")
lines := reload(true)
require.Len(t, lines, 2)
assert.Contains(t, lines[1], "Config reload failed, the new config did not load")
assert.Contains(t, lines[1], dir, "names the config that did not load")
}
// A client that went away still gets the reload it asked for
func TestReloadCommandReloadsWhenTheClientIsGone(t *testing.T) {
dir := t.TempDir()
require.NoError(t, os.WriteFile(filepath.Join(dir, "config.yml"), []byte("a: 1"), 0o600))
l := test.NewLogger()
c := config.NewC(l)
require.NoError(t, c.Load(dir))
reloaded := false
c.RegisterReloadCallback(func(*config.C, *config.Reload) { reloaded = true })
gone := errors.New("client is gone")
require.ErrorIs(t, reloadCommand(l, c, &lineWriter{err: gone}), gone)
assert.True(t, reloaded)
}
-161
View File
@@ -1,161 +0,0 @@
package sshd
import (
"errors"
"flag"
"fmt"
"sort"
"strings"
"github.com/armon/go-radix"
)
// CommandFlags is a function called before help or command execution to parse command line flags
// It should return a flag.FlagSet instance and a pointer to the struct that will contain parsed flags
type CommandFlags func() (*flag.FlagSet, any)
// CommandCallback is the function called when your command should execute.
// fs will be a a pointer to the struct provided by Command.Flags callback, if there was one. -h and -help are reserved
// and handled automatically for you.
// a will be any unconsumed arguments, if no Command.Flags was available this will be all the flags passed in.
// w is the writer to use when sending messages back to the client.
// If an error is returned by the callback it is logged locally, the callback should handle messaging errors to the user
// where appropriate
type CommandCallback func(fs any, a []string, w StringWriter) error
type Command struct {
Name string
ShortDescription string
Help string
Flags CommandFlags
Callback CommandCallback
}
func execCommand(c *Command, args []string, w StringWriter) error {
var (
fl *flag.FlagSet
fs any
)
if c.Flags != nil {
fl, fs = c.Flags()
if fl != nil {
// SetOutput() here in case fl.Parse dumps usage.
fl.SetOutput(w.GetWriter())
err := fl.Parse(args)
if err != nil {
// fl.Parse has dumped error information to the user via the w writer.
return err
}
args = fl.Args()
}
}
return c.Callback(fs, args, w)
}
func dumpCommands(c *radix.Tree, w StringWriter) {
err := w.WriteLine("Available commands:")
if err != nil {
return
}
cmds := make([]string, 0)
for _, l := range allCommands(c) {
cmds = append(cmds, fmt.Sprintf("%s - %s", l.Name, l.ShortDescription))
}
sort.Strings(cmds)
_ = w.Write(strings.Join(cmds, "\n") + "\n\n")
}
func lookupCommand(c *radix.Tree, sCmd string) (*Command, error) {
cmd, ok := c.Get(sCmd)
if !ok {
return nil, nil
}
command, ok := cmd.(*Command)
if !ok {
return nil, errors.New("failed to cast command")
}
return command, nil
}
func matchCommand(c *radix.Tree, cmd string) []string {
cmds := make([]string, 0)
c.WalkPrefix(cmd, func(found string, v any) bool {
cmds = append(cmds, found)
return false
})
sort.Strings(cmds)
return cmds
}
func allCommands(c *radix.Tree) []*Command {
cmds := make([]*Command, 0)
c.WalkPrefix("", func(found string, v any) bool {
cmd, ok := v.(*Command)
if ok {
cmds = append(cmds, cmd)
}
return false
})
return cmds
}
func helpCallback(commands *radix.Tree, a []string, w StringWriter) (err error) {
// Just typed help
if len(a) == 0 {
dumpCommands(commands, w)
return nil
}
// We are printing a specific commands help text
cmd, err := lookupCommand(commands, a[0])
if err != nil {
return
}
if cmd != nil {
err = w.WriteLine(fmt.Sprintf("%s - %s", cmd.Name, cmd.ShortDescription))
if err != nil {
return err
}
if cmd.Help != "" {
err = w.WriteLine(fmt.Sprintf(" %s", cmd.Help))
if err != nil {
return err
}
}
if cmd.Flags != nil {
fs, _ := cmd.Flags()
if fs != nil {
fs.SetOutput(w.GetWriter())
fs.PrintDefaults()
}
}
return nil
}
err = w.WriteLine("Command not available " + a[0])
if err != nil {
return err
}
return nil
}
func checkHelpArgs(args []string) bool {
for _, a := range args {
if a == "-h" || a == "-help" {
return true
}
}
return false
}
+75 -41
View File
@@ -8,13 +8,17 @@ import (
"log/slog"
"net"
"sync"
"sync/atomic"
"github.com/armon/go-radix"
"github.com/slackhq/nebula/commands"
"golang.org/x/crypto/ssh"
)
type SSHServer struct {
config *ssh.ServerConfig
// config is replaced whole when the host key changes, ssh.ServerConfig.AddHostKey only replaces a key of the same
// algorithm and appends any other
config atomic.Pointer[ssh.ServerConfig]
authenticate func(ssh.ConnMetadata, ssh.PublicKey) (*ssh.Permissions, error)
l *slog.Logger
certChecker *ssh.CertChecker
@@ -25,24 +29,28 @@ type SSHServer struct {
trustedKeys map[string]map[string]bool
trustedCAs []ssh.PublicKey
// List of available commands
helpCommand *Command
commands *radix.Tree
listener net.Listener
// Available commands.
commands *commands.Registry
listener atomic.Pointer[net.Listener]
// runs counts the Runs that are listening. A reload's Run can start before the one it replaced has returned
runs atomic.Int32
// addr is where the newest Run listens, nil when none is
addr atomic.Pointer[net.Addr]
// ctx parents per-Run contexts. Cancelling it (e.g. via Control.Stop) tears the server down even
// across reloads, since each Run derives a fresh child rather than reusing this one directly.
ctx context.Context
}
// NewSSHServer creates a new ssh server rigged with default commands and prepares to listen.
// The ssh server's context is parented off the supplied ctx so cancelling it
// (e.g. on Control.Stop) tears down active sessions and closes the listener.
func NewSSHServer(ctx context.Context, l *slog.Logger) (*SSHServer, error) {
// NewSSHServer creates a new ssh server serving the supplied command registry and
// prepares to listen. The ssh server's context is parented off the supplied ctx so
// cancelling it (e.g. on Control.Stop) tears down active sessions and closes the listener.
func NewSSHServer(ctx context.Context, l *slog.Logger, reg *commands.Registry) (*SSHServer, error) {
s := &SSHServer{
trustedKeys: make(map[string]map[string]bool),
l: l,
commands: radix.New(),
commands: reg,
ctx: ctx,
}
@@ -85,18 +93,8 @@ func NewSSHServer(ctx context.Context, l *slog.Logger) (*SSHServer, error) {
},
}
s.config = &ssh.ServerConfig{
PublicKeyCallback: cc.Authenticate,
ServerVersion: fmt.Sprintf("SSH-2.0-Nebula???"),
}
s.RegisterCommand(&Command{
Name: "help",
ShortDescription: "prints available commands or help <command> for specific usage info",
Callback: func(a any, args []string, w StringWriter) error {
return helpCallback(s.commands, args, w)
},
})
s.authenticate = cc.Authenticate
s.config.Store(s.newConfig())
return s, nil
}
@@ -107,10 +105,32 @@ func (s *SSHServer) SetHostKey(hostPrivateKey []byte) error {
return fmt.Errorf("failed to parse private key: %s", err)
}
s.config.AddHostKey(private)
config := s.newConfig()
config.AddHostKey(private)
s.config.Store(config)
return nil
}
func (s *SSHServer) newConfig() *ssh.ServerConfig {
return &ssh.ServerConfig{
PublicKeyCallback: s.authenticate,
ServerVersion: fmt.Sprintf("SSH-2.0-Nebula???"),
}
}
// Addr is the address the newest Run listens on, nil when none is.
func (s *SSHServer) Addr() net.Addr {
if a := s.addr.Load(); a != nil {
return *a
}
return nil
}
// Running is true while a Run is listening.
func (s *SSHServer) Running() bool {
return s.runs.Load() > 0
}
func (s *SSHServer) ClearTrustedCAs() {
s.authLock.Lock()
s.trustedCAs = []ssh.PublicKey{}
@@ -160,31 +180,47 @@ func (s *SSHServer) AddAuthorizedKey(user, pubKey string) error {
return nil
}
// RegisterCommand adds a command that can be run by a user, by default only `help` is available
func (s *SSHServer) RegisterCommand(c *Command) {
s.commands.Insert(c.Name, c)
}
// Run begins listening and accepting connections. Each invocation derives a fresh per-Run context
// from the constructor-supplied ctx so a Stop+Run sequence (used by config reload) starts clean
// rather than carrying a permanently-cancelled context across runs.
func (s *SSHServer) Run(addr string) error {
if s.ctx.Err() != nil {
return s.ctx.Err()
}
listener, err := net.Listen("tcp", addr)
listener, err := s.Listen(addr)
if err != nil {
return err
}
s.Serve(listener)
return nil
}
// Listen binds addr for Serve, so a caller can tell the bind failed before serving in the background. Stop can close
// the listener from here on, Serve then returns as soon as it starts.
func (s *SSHServer) Listen(addr string) (net.Listener, error) {
if s.ctx.Err() != nil {
return nil, s.ctx.Err()
}
listener, err := net.Listen("tcp", addr)
if err != nil {
return nil, err
}
// s.listener is the public handle Stop uses to interrupt the active run; listener (the local) is what
// this run owns. They start equal but a fast reload may overwrite s.listener with the next run's
// listener before this run's watcher fires, so each run must close its own listener via the local
// reference.
s.listener = listener
s.listener.Store(&listener)
return listener, nil
}
// Serve accepts connections on a listener from Listen until Stop, or the context passed to NewSSHServer, ends it.
func (s *SSHServer) Serve(listener net.Listener) {
s.runs.Add(1)
bound := listener.Addr()
s.addr.Store(&bound)
runCtx, cancel := context.WithCancel(s.ctx)
defer cancel()
// Deferred after cancel so they run first, a run on its way out no longer counts as listening
defer s.runs.Add(-1)
defer s.addr.CompareAndSwap(&bound, nil)
// Close the listener when this run's context is cancelled. That can come from the parent
// (Control.Stop), from Run returning normally (defer cancel above), or transitively when a sibling
@@ -197,14 +233,12 @@ func (s *SSHServer) Run(addr string) error {
}
}()
s.l.Info("SSH server is listening", "sshListener", addr)
s.l.Info("SSH server is listening", "sshListener", bound)
// Run loops until there is an error
s.run(runCtx, listener)
s.l.Info("SSH server stopped listening")
// We don't return an error because run logs for us
return nil
}
func (s *SSHServer) run(ctx context.Context, listener net.Listener) {
@@ -226,7 +260,7 @@ func (s *SSHServer) run(ctx context.Context, listener net.Listener) {
<-sessionContext.Done()
c.Close()
}()
conn, chans, reqs, err := ssh.NewServerConn(c, s.config)
conn, chans, reqs, err := ssh.NewServerConn(c, s.config.Load())
fp := ""
if conn != nil {
fp = conn.Permissions.Extensions["fp"]
@@ -264,8 +298,8 @@ func (s *SSHServer) run(ctx context.Context, listener net.Listener) {
}
func (s *SSHServer) Stop() {
if s.listener != nil {
if err := s.listener.Close(); err != nil {
if l := s.listener.Load(); l != nil {
if err := (*l).Close(); err != nil && !errors.Is(err, net.ErrClosed) {
s.l.Warn("Failed to close the sshd listener", "error", err)
}
}
+37 -46
View File
@@ -1,13 +1,11 @@
package sshd
import (
"fmt"
"errors"
"log/slog"
"sort"
"strings"
"github.com/anmitsu/go-shlex"
"github.com/armon/go-radix"
"github.com/slackhq/nebula/commands"
"golang.org/x/crypto/ssh"
"golang.org/x/term"
)
@@ -16,31 +14,37 @@ type session struct {
l *slog.Logger
c *ssh.ServerConn
term *term.Terminal
commands *radix.Tree
commands *commands.Registry
cancel func()
}
func NewSession(commands *radix.Tree, conn *ssh.ServerConn, chans <-chan ssh.NewChannel, cancel func(), l *slog.Logger) *session {
func NewSession(reg *commands.Registry, conn *ssh.ServerConn, chans <-chan ssh.NewChannel, cancel func(), l *slog.Logger) *session {
s := &session{
commands: radix.NewFromMap(commands.ToMap()),
commands: reg.Clone(),
l: l,
c: conn,
cancel: cancel,
}
s.commands.Insert("logout", &Command{
Name: "logout",
ShortDescription: "Ends the current session",
Callback: func(a any, args []string, w StringWriter) error {
s.Close()
return nil
},
})
s.commands.Register(LogoutCommand(s.Close))
go s.handleChannels(chans)
return s
}
// LogoutCommand is the logout command a session registers on its clone of the registry,
// with end as what it does. Exported so a test can serve what a session serves.
func LogoutCommand(end func()) *commands.Command {
return &commands.Command{
Name: commands.Logout,
ShortDescription: "Ends the current session",
Callback: func(fs any, a []string, w commands.StringWriter) error {
end()
return nil
},
}
}
func (s *session) handleChannels(chans <-chan ssh.NewChannel) {
defer s.Close()
for newChannel := range chans {
@@ -87,9 +91,13 @@ func (s *session) handleRequests(in <-chan *ssh.Request, channel ssh.Channel) {
}
req.Reply(true, nil)
s.dispatchCommand(payload.Value, &stringWriter{channel})
err := s.dispatchCommand(payload.Value, commands.NewStringWriter(channel))
// A command that failed exits 1, so `ssh host cmd` callers can tell.
status := struct{ Status uint32 }{uint32(0)}
if err != nil {
status.Status = 1
}
channel.SendRequest("exit-status", false, ssh.Marshal(status))
channel.Close()
return
@@ -111,12 +119,11 @@ func (s *session) createTerm(channel ssh.Channel) *term.Terminal {
term.AutoCompleteCallback = func(line string, pos int, key rune) (newLine string, newPos int, ok bool) {
// key 9 is tab
if key == 9 {
cmds := matchCommand(s.commands, line)
cmds := s.commands.Match(line)
if len(cmds) == 1 {
return cmds[0] + " ", len(cmds[0]) + 1, true
}
sort.Strings(cmds)
term.Write([]byte(strings.Join(cmds, "\n") + "\n\n"))
}
@@ -128,7 +135,7 @@ func (s *session) createTerm(channel ssh.Channel) *term.Terminal {
}
func (s *session) handleInput() {
w := &stringWriter{w: s.term}
w := commands.NewStringWriter(s.term)
for {
line, err := s.term.ReadLine()
if err != nil {
@@ -139,36 +146,20 @@ func (s *session) handleInput() {
}
}
func (s *session) dispatchCommand(line string, w StringWriter) {
args, err := shlex.Split(line, true)
if err != nil {
return
// dispatchCommand runs line and returns the command's error, which it has logged unless
// the command already told the user everything: the log line is for the operator reading
// the daemon's log, who learns nothing from a failure the user caused.
func (s *session) dispatchCommand(line string, w commands.StringWriter) error {
err := s.commands.Dispatch(line, w)
if err == nil {
return nil
}
if len(args) == 0 {
dumpCommands(s.commands, w)
return
var f *commands.Failure
if !errors.As(err, &f) {
s.l.Debug("Command returned an error", "line", line, "error", err)
}
c, err := lookupCommand(s.commands, args[0])
if err != nil {
return
}
if c == nil {
err := w.WriteLine(fmt.Sprintf("did not understand: %s", line))
_ = err
dumpCommands(s.commands, w)
return
}
if checkHelpArgs(args) {
s.dispatchCommand(fmt.Sprintf("%s %s", "help", c.Name), w)
return
}
_ = execCommand(c, args[1:], w)
return err
}
func (s *session) Close() {
+293
View File
@@ -0,0 +1,293 @@
package nebula
import (
"crypto"
"crypto/ed25519"
"crypto/rand"
"crypto/rsa"
"encoding/pem"
"errors"
"fmt"
"net"
"os"
"path/filepath"
"strings"
"testing"
"time"
"github.com/slackhq/nebula/commands"
"github.com/slackhq/nebula/config"
"github.com/slackhq/nebula/sshd"
"github.com/slackhq/nebula/test"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
"golang.org/x/crypto/ssh"
)
type sshdReloadHarness struct {
c *config.C
reg *commands.Registry
server *sshd.SSHServer
keyFile string
signer ssh.Signer
listen string
users []string
extra string
}
func newSSHDReloadHarness(t *testing.T) *sshdReloadHarness {
t.Helper()
l := test.NewLogger()
h := &sshdReloadHarness{keyFile: filepath.Join(t.TempDir(), "host_key"), listen: "127.0.0.1:0", users: []string{"test"}}
h.writeHostKey(t, newEd25519Key(t))
_, userKey, err := ed25519.GenerateKey(rand.Reader)
require.NoError(t, err)
h.signer, err = ssh.NewSignerFromKey(userKey)
require.NoError(t, err)
h.c = config.NewC(l)
require.NoError(t, h.c.LoadString(h.config()))
h.reg = commands.New()
h.server, err = sshd.NewSSHServer(t.Context(), l, h.reg)
require.NoError(t, err)
wireSSHReload(l, h.server, h.c)
run, err := configSSH(l, h.server, h.c, nil)
require.NoError(t, err)
go run()
t.Cleanup(h.server.Stop)
require.Eventually(t, h.server.Running, 5*time.Second, 10*time.Millisecond)
return h
}
func newEd25519Key(t *testing.T) crypto.Signer {
_, key, err := ed25519.GenerateKey(rand.Reader)
require.NoError(t, err)
return key
}
// writeHostKey replaces the host key file and returns the public key a client should now see
func (h *sshdReloadHarness) writeHostKey(t *testing.T, key crypto.Signer) ssh.PublicKey {
t.Helper()
block, err := ssh.MarshalPrivateKey(key, "")
require.NoError(t, err)
require.NoError(t, os.WriteFile(h.keyFile, pem.EncodeToMemory(block), 0o600))
pub, err := ssh.NewPublicKey(key.Public())
require.NoError(t, err)
return pub
}
func (h *sshdReloadHarness) config() string {
users := ""
for _, u := range h.users {
users += fmt.Sprintf(" - user: %s\n keys: [%q]\n", u, string(ssh.MarshalAuthorizedKey(h.signer.PublicKey())))
}
return fmt.Sprintf("sshd:\n enabled: true\n listen: %s\n host_key: %s\n authorized_users:\n%s%s",
h.listen, h.keyFile, users, h.extra)
}
// dial connects to wherever the server listens now, a restart on :0 moves it. hostKeyAlgorithms limits which host keys
// the client accepts, none means any
func (h *sshdReloadHarness) dial(hostKeyAlgorithms ...string) (*ssh.Client, ssh.PublicKey, error) {
addr := h.server.Addr()
if addr == nil {
return nil, nil, fmt.Errorf("not listening")
}
var hostKey ssh.PublicKey
c, err := ssh.Dial("tcp", addr.String(), &ssh.ClientConfig{
User: "test",
Auth: []ssh.AuthMethod{ssh.PublicKeys(h.signer)},
HostKeyCallback: func(_ string, _ net.Addr, key ssh.PublicKey) error {
hostKey = key
return nil
},
HostKeyAlgorithms: hostKeyAlgorithms,
Timeout: time.Second,
})
return c, hostKey, err
}
func (h *sshdReloadHarness) connect(t *testing.T) (*ssh.Client, ssh.PublicKey) {
t.Helper()
var client *ssh.Client
var hostKey ssh.PublicKey
require.Eventually(t, func() bool {
var err error
client, hostKey, err = h.dial()
return err == nil
}, 5*time.Second, 20*time.Millisecond)
t.Cleanup(func() { _ = client.Close() })
return client, hostKey
}
// dropped is true when the server closed the client's connection within wait
func dropped(client *ssh.Client, wait time.Duration) bool {
done := make(chan struct{})
go func() {
_ = client.Wait()
close(done)
}()
select {
case <-done:
return true
case <-time.After(wait):
return false
}
}
// Restarting sshd drops every open session, the one that asked for the reload included. A reload that touches nothing
// sshd reads leaves it alone
func TestSSHDReload_UnchangedKeepsSessions(t *testing.T) {
h := newSSHDReloadHarness(t)
client, _ := h.connect(t)
require.NoError(t, h.c.ReloadConfigString(h.config()+"punchy:\n punch: true\n").Err)
assert.False(t, dropped(client, 500*time.Millisecond), "an unchanged sshd was restarted")
}
// Every key the server reads restarts it when it changes
func TestSSHDReload_ChangedConfigRestarts(t *testing.T) {
cases := map[string]func(t *testing.T, h *sshdReloadHarness){
"listen": func(t *testing.T, h *sshdReloadHarness) { h.listen = "localhost:0" },
"authorized_users": func(t *testing.T, h *sshdReloadHarness) { h.users = append(h.users, "other") },
"trusted_cas": func(t *testing.T, h *sshdReloadHarness) { h.extra = " trusted_cas: []\n" },
"host_key": func(t *testing.T, h *sshdReloadHarness) {
h.keyFile = filepath.Join(t.TempDir(), "other_host_key")
h.writeHostKey(t, newEd25519Key(t))
},
}
for name, change := range cases {
t.Run(name, func(t *testing.T) {
h := newSSHDReloadHarness(t)
client, _ := h.connect(t)
change(t, h)
require.NoError(t, h.c.ReloadConfigString(h.config()).Err)
assert.True(t, dropped(client, 5*time.Second), "a changed sshd config was not applied")
require.Eventually(t, h.server.Running, 5*time.Second, 10*time.Millisecond)
h.connect(t)
})
}
}
// The config can stay the same while the host key file it names changes. New connections get the new key, open
// sessions are left alone
func TestSSHDReload_ChangedHostKeyFile(t *testing.T) {
h := newSSHDReloadHarness(t)
client, _ := h.connect(t)
want := h.writeHostKey(t, newEd25519Key(t))
require.NoError(t, h.c.ReloadConfigString(h.config()).Err)
_, got := h.connect(t)
assert.Equal(t, want.Marshal(), got.Marshal(), "a new host key was not applied")
assert.False(t, dropped(client, 500*time.Millisecond), "a new host key restarted sshd")
}
// A host key of another algorithm replaces the old one rather than being offered next to it
func TestSSHDReload_HostKeyAlgorithmChange(t *testing.T) {
h := newSSHDReloadHarness(t)
h.connect(t)
rsaKey, err := rsa.GenerateKey(rand.Reader, 2048)
require.NoError(t, err)
want := h.writeHostKey(t, rsaKey)
require.NoError(t, h.c.ReloadConfigString(h.config()).Err)
_, got := h.connect(t)
assert.Equal(t, want.Marshal(), got.Marshal())
_, _, err = h.dial(ssh.KeyAlgoED25519)
assert.Error(t, err, "the old ed25519 host key is still offered")
}
// A server that isn't running, say its port was busy, is tried again on every reload as before
func TestSSHDReload_NotRunningRestarts(t *testing.T) {
h := newSSHDReloadHarness(t)
h.server.Stop()
require.Eventually(t, func() bool { return !h.server.Running() }, 5*time.Second, 10*time.Millisecond)
require.NoError(t, h.c.ReloadConfigString(h.config()).Err)
require.Eventually(t, h.server.Running, 5*time.Second, 10*time.Millisecond)
h.connect(t)
}
// A host key that no longer loads stops sshd, the same as a config that doesn't
func TestSSHDReload_BadHostKeyFileStops(t *testing.T) {
h := newSSHDReloadHarness(t)
h.connect(t)
require.NoError(t, os.WriteFile(h.keyFile, []byte("not a key"), 0o600))
assert.True(t, h.c.ReloadConfigString(h.config()).Failed)
require.Eventually(t, func() bool { return !h.server.Running() }, 5*time.Second, 10*time.Millisecond)
}
// A trusted CA or authorized user that is skipped with a warning fails the reload, the valid ones are still used
func TestSSHDReload_SkippedEntriesFail(t *testing.T) {
cases := map[string]string{
"bad trusted ca": " trusted_cas: [nope]\n",
"user entry not a map": " - nope\n",
"user missing": " - keys: [nope]\n",
"bad key": " - user: other\n keys: nope\n",
"key in a list not a string": " - user: other\n keys: [5]\n",
"bad key in a list": " - user: other\n keys: [nope]\n",
"keys missing": " - user: other\n",
}
for name, extra := range cases {
t.Run(name, func(t *testing.T) {
h := newSSHDReloadHarness(t)
h.extra = extra
res := h.c.ReloadConfigString(h.config())
require.NoError(t, res.Err)
assert.True(t, res.Failed)
require.Eventually(t, h.server.Running, 5*time.Second, 10*time.Millisecond)
h.connect(t)
})
}
}
// TestSSHExecExitStatusReflectsTheCommand: an exec request exits 1 when the command
// returned an error and 0 otherwise, with the command's output either way.
func TestSSHExecExitStatusReflectsTheCommand(t *testing.T) {
h := newSSHDReloadHarness(t)
h.reg.Register(&commands.Command{
Name: "fails",
Callback: func(fs any, a []string, w commands.StringWriter) error {
_ = w.WriteLine("nope")
return errors.New("failed")
},
})
client, _ := h.connect(t)
sess, err := client.NewSession()
require.NoError(t, err)
out, err := sess.CombinedOutput("fails")
var exit *ssh.ExitError
require.ErrorAs(t, err, &exit)
assert.Equal(t, 1, exit.ExitStatus())
assert.Equal(t, "nope\n", string(out))
sess, err = client.NewSession()
require.NoError(t, err)
out, err = sess.CombinedOutput("help fails")
require.NoError(t, err)
assert.Equal(t, "fails - \n", string(out))
// A word that is not a command is a failure too, with the command list as before.
sess, err = client.NewSession()
require.NoError(t, err)
out, err = sess.CombinedOutput("nope")
require.ErrorAs(t, err, &exit)
assert.Equal(t, 1, exit.ExitStatus())
assert.True(t, strings.HasPrefix(string(out), "did not understand: nope\nAvailable commands:\n"), string(out))
}
// A reload that can't bind the new address says so, rather than finding out in the background
func TestSSHDReload_BusyPortFails(t *testing.T) {
h := newSSHDReloadHarness(t)
busy, err := net.Listen("tcp", "127.0.0.1:0")
require.NoError(t, err)
t.Cleanup(func() { _ = busy.Close() })
res := h.c.ReloadConfigString(strings.Replace(h.config(), "127.0.0.1:0", busy.Addr().String(), 1))
require.NoError(t, res.Err)
assert.True(t, res.Failed)
}
+88 -31
View File
@@ -38,6 +38,11 @@ type statsServer struct {
runMu sync.Mutex
runCfg *statsConfig
run *statsRuntime // non-nil while a runtime is live
// Closed when the last runtime's capture loop has returned. Stop waits on it so nothing captures once it returns,
// and a reload's new runtime never overlaps the old one, go-metrics keeps its GC capture state in unguarded globals
loopDone chan struct{}
// See newStatsServerFromConfig, never changes after it
emitStats func()
}
// statsRuntime is the live state owned by a single Start invocation. Start
@@ -82,17 +87,22 @@ type promConfig struct {
//
// Start is safe to call unconditionally: it no-ops when stats are disabled.
// The returned pointer is always non-nil, even on error.
func newStatsServerFromConfig(ctx context.Context, l *slog.Logger, c *config.C, buildVersion string, configTest bool) (*statsServer, error) {
//
// emitStats, if set, updates metrics on every capture, ahead of the exporters so each export sees what it just set. It
// is fixed here, before the reload callback exists, so every runtime has it.
func newStatsServerFromConfig(ctx context.Context, l *slog.Logger, c *config.C, buildVersion string, configTest bool, emitStats func()) (*statsServer, error) {
s := &statsServer{
l: l,
ctx: ctx,
buildVersion: buildVersion,
configTest: configTest,
emitStats: emitStats,
}
c.RegisterReloadCallback(func(c *config.C) {
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
if err := s.reload(c, false); err != nil {
s.l.Error("Failed to reload stats from config", "error", err)
r.Failed()
}
})
@@ -120,7 +130,6 @@ func (s *statsServer) reload(c *config.C, initial bool) error {
s.runMu.Lock()
sameCfg := s.runCfg != nil && *s.runCfg == newCfg
s.runCfg = &newCfg
running := s.run != nil
s.runMu.Unlock()
s.enabled.Store(enabled)
@@ -129,11 +138,15 @@ func (s *statsServer) reload(c *config.C, initial bool) error {
return nil
}
if running {
// Even with nothing running, a runtime whose listener failed can still be finishing its last pass
s.Stop()
// Bind here rather than in the background, so a reload that can't listen says so
run, err := s.start()
if err != nil {
return err
}
if enabled && !s.configTest {
go s.Start()
if run != nil {
go run()
}
return nil
}
@@ -143,30 +156,64 @@ func (s *statsServer) reload(c *config.C, initial bool) error {
// the HTTP listener. For graphite it blocks on the capture loop's context.
// Safe to call when stats are disabled or already running (both no-op).
func (s *statsServer) Start() {
if !s.enabled.Load() || s.configTest {
run, err := s.start()
if err != nil {
s.l.Error("Failed to start the prometheus stats listener", "error", err)
return
}
if run != nil {
run()
}
}
// start claims the runtime and binds the prometheus listener. The returned run serves it and blocks until Stop or ctx
// ends it. A nil run and no error means stats are off or already running.
func (s *statsServer) start() (func(), error) {
if !s.enabled.Load() || s.configTest {
return nil, nil
}
s.runMu.Lock()
defer s.runMu.Unlock()
if s.ctx.Err() != nil || s.run != nil || s.runCfg == nil {
s.runMu.Unlock()
return
return nil, nil
}
cfg := *s.runCfg
captureFns, listener := s.buildRuntime(cfg)
runCtx, cancel := context.WithCancel(s.ctx)
captureFns, listener := s.buildRuntime(runCtx, cfg)
var ln net.Listener
if listener != nil {
var err error
ln, err = net.Listen("tcp", listener.Addr)
if err != nil {
cancel()
// Drop the config so a SIGHUP with the same one tries again once the problem is fixed
s.runCfg = nil
return nil, err
}
}
rt := &statsRuntime{cancel: cancel, listener: listener}
s.run = rt
s.runMu.Unlock()
done := make(chan struct{})
s.loopDone = done
go captureStatsLoop(runCtx, cfg.interval, captureFns)
return func() { s.serve(runCtx, rt, captureFns, cfg.interval, ln, done) }, nil
}
// serve runs the capture loop, and the prometheus listener on ln, until Stop or ctx ends them.
func (s *statsServer) serve(runCtx context.Context, rt *statsRuntime, captureFns []func(), interval time.Duration, ln net.Listener, done chan struct{}) {
// Prime before serving, so a scrape that lands before the first tick sees real values instead of zeros (issue #907)
for _, fn := range captureFns {
fn()
}
go captureStatsLoop(runCtx, interval, captureFns, done)
cleanExit := true
if listener == nil {
if rt.listener == nil {
// Graphite: no HTTP listener to serve; block until teardown.
<-runCtx.Done()
} else {
cleanExit = s.serveListener(listener)
cleanExit = s.serveListener(rt.listener, ln)
}
// Clear our runtime only if nothing has replaced it. Stop races through
@@ -175,10 +222,9 @@ func (s *statsServer) Start() {
if s.run == rt {
rt.cancel()
s.run = nil
// A listener that exited with an error (e.g., bind conflict) leaves
// runCfg cached as if it were applied. Drop it so a SIGHUP with the
// same config re-triggers Start once the user fixes the underlying
// problem.
// A listener that exited with an error leaves runCfg cached as if it
// were applied. Drop it so a SIGHUP with the same config re-triggers
// Start once the user fixes the underlying problem.
if !cleanExit {
s.runCfg = nil
}
@@ -186,12 +232,12 @@ func (s *statsServer) Start() {
s.runMu.Unlock()
}
// serveListener runs ListenAndServe and ensures ctx cancellation unblocks it.
// serveListener serves listener on ln and ensures ctx cancellation unblocks it.
// Returns true if the listener exited cleanly (Stop, ctx cancellation, or any
// other http.ErrServerClosed path), false on an unexpected error.
func (s *statsServer) serveListener(listener *http.Server) bool {
func (s *statsServer) serveListener(listener *http.Server, ln net.Listener) bool {
// Per-invocation watcher: ctx cancellation triggers a listener shutdown
// which in turn unblocks ListenAndServe. Closing `done` on exit keeps
// which in turn unblocks Serve. Closing `done` on exit keeps
// the watcher from outliving this call.
done := make(chan struct{})
go func() {
@@ -207,8 +253,8 @@ func (s *statsServer) serveListener(listener *http.Server) bool {
}()
defer close(done)
s.l.Info("Starting prometheus stats listener", "addr", listener.Addr)
err := listener.ListenAndServe()
s.l.Info("Starting prometheus stats listener", "addr", ln.Addr())
err := listener.Serve(ln)
if err == nil || errors.Is(err, http.ErrServerClosed) {
return true
}
@@ -218,13 +264,12 @@ func (s *statsServer) serveListener(listener *http.Server) bool {
// Stop tears down the active runtime, if any. Idempotent.
func (s *statsServer) Stop() {
// Held until the last pass is done, a Start that slipped in after s.run cleared would capture alongside it
s.runMu.Lock()
defer s.runMu.Unlock()
rt := s.run
s.run = nil
s.runMu.Unlock()
if rt == nil {
return
}
if rt != nil {
rt.cancel()
if rt.listener != nil {
shutdownCtx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
@@ -234,10 +279,17 @@ func (s *statsServer) Stop() {
cancel()
}
}
// Nothing captures once Stop returns. The last runtime is cancelled by now, whoever cleared it, so this is at most
// the pass it was in. Nothing in a pass takes runMu
if s.loopDone != nil {
<-s.loopDone
}
}
// buildRuntime produces the capture functions and, for prometheus, an un-served
// http.Server from cfg. cfg has already been validated by loadStatsConfig.
func (s *statsServer) buildRuntime(cfg statsConfig) ([]func(), *http.Server) {
// A graphite send in progress is abandoned when ctx is done.
func (s *statsServer) buildRuntime(ctx context.Context, cfg statsConfig) ([]func(), *http.Server) {
// rcrowley/go-metrics guards these registrations with a private sync.Once,
// so subsequent reloads are no-ops.
metrics.RegisterDebugGCStats(metrics.DefaultRegistry)
@@ -247,6 +299,9 @@ func (s *statsServer) buildRuntime(cfg statsConfig) ([]func(), *http.Server) {
func() { metrics.CaptureDebugGCStatsOnce(metrics.DefaultRegistry) },
func() { metrics.CaptureRuntimeMemStatsOnce(metrics.DefaultRegistry) },
}
if s.emitStats != nil {
captureFns = append(captureFns, s.emitStats)
}
switch cfg.typ {
case "graphite":
@@ -262,7 +317,7 @@ func (s *statsServer) buildRuntime(cfg statsConfig) ([]func(), *http.Server) {
Percentiles: []float64{0.5, 0.75, 0.95, 0.99, 0.999},
}
captureFns = append(captureFns, func() {
if err := graphiteOnce(gcfg); err != nil {
if err := graphiteOnce(ctx, gcfg, graphiteTimeout); err != nil && ctx.Err() == nil {
s.l.Error("Graphite export failed", "error", err)
}
})
@@ -309,8 +364,10 @@ func (s *statsServer) buildRuntime(cfg statsConfig) ([]func(), *http.Server) {
return captureFns, nil
}
// captureStatsLoop runs each fn on every tick of d until ctx is cancelled.
func captureStatsLoop(ctx context.Context, d time.Duration, fns []func()) {
// captureStatsLoop runs each fn on every tick of d until ctx is cancelled, then
// closes done.
func captureStatsLoop(ctx context.Context, d time.Duration, fns []func(), done chan struct{}) {
defer close(done)
t := time.NewTicker(d)
defer t.Stop()
for {
+319 -29
View File
@@ -2,13 +2,18 @@ package nebula
import (
"context"
"fmt"
"io"
"log/slog"
"net"
"net/http"
"net/http/httptest"
"strconv"
"sync/atomic"
"testing"
"time"
"github.com/rcrowley/go-metrics"
"github.com/slackhq/nebula/config"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
@@ -110,15 +115,14 @@ func TestStatsServer_reload_initial_prometheus(t *testing.T) {
}
func TestStatsServer_Start_graphite_blocksUntilStop(t *testing.T) {
sink := newGraphiteSink(t)
defer sink.Close()
h := newGraphiteHost(t, 0)
s, c := newTestStatsServer(t)
setStatsConfig(c, map[string]any{
"type": "graphite",
"interval": "1s",
"protocol": "tcp",
"host": sink.Addr(),
"host": h.ln.Addr().String(),
"prefix": "test",
})
require.NoError(t, s.reload(c, true))
@@ -371,34 +375,18 @@ func waitForListening(t *testing.T, addr string) {
})
}
// graphiteSink is a minimal TCP accept-and-discard server so graphiteOnce
// calls in tests don't spam error logs or wedge on connection refused.
type graphiteSink struct {
ln net.Listener
// stopAtCleanup returns the ctx to build *s with, and stops it when the test ends. Cancelling first turns away a Start
// a reload spawned that hasn't run yet, then Stop waits out the last pass, so nothing captures into the next test
func stopAtCleanup(t *testing.T, s **statsServer) context.Context {
ctx, cancel := context.WithCancel(context.Background())
t.Cleanup(func() {
cancel()
if *s != nil {
(*s).Stop()
}
func newGraphiteSink(t *testing.T) *graphiteSink {
t.Helper()
ln, err := net.Listen("tcp", "127.0.0.1:0")
require.NoError(t, err)
g := &graphiteSink{ln: ln}
go func() {
for {
conn, err := ln.Accept()
if err != nil {
return
})
return ctx
}
go func(c net.Conn) {
_, _ = io.Copy(io.Discard, c)
_ = c.Close()
}(conn)
}
}()
return g
}
func (g *graphiteSink) Addr() string { return g.ln.Addr().String() }
func (g *graphiteSink) Close() { _ = g.ln.Close() }
func freeTCPPort(t *testing.T) string {
t.Helper()
@@ -408,3 +396,305 @@ func freeTCPPort(t *testing.T) string {
require.NoError(t, ln.Close())
return strconv.Itoa(port)
}
// Each export has to see what emitStats set in the same pass
func TestStatsServer_emitStatsRunsBeforeExport(t *testing.T) {
defer metrics.DefaultRegistry.Unregister("emitter.order.test")
l := slog.New(slog.DiscardHandler)
c := config.NewC(l)
setStatsConfig(c, map[string]any{
"type": "prometheus",
"interval": "1s",
"listen": "127.0.0.1:0",
"path": "/metrics",
})
gauge := metrics.GetOrRegisterGauge("emitter.order.test", nil)
s, err := newStatsServerFromConfig(t.Context(), l, c, "", false, func() { gauge.Update(42) })
require.NoError(t, err)
s.runMu.Lock()
fns, srv := s.buildRuntime(t.Context(), *s.runCfg)
s.runMu.Unlock()
require.NotNil(t, srv)
for _, fn := range fns {
fn()
}
rec := httptest.NewRecorder()
srv.Handler.ServeHTTP(rec, httptest.NewRequest("GET", "/metrics", nil))
assert.Contains(t, rec.Body.String(), "emitter_order_test 42")
}
// A scrape that lands the moment the listener is up already sees what emitStats set (issue #907)
func TestStatsServer_Start_primesBeforeServing(t *testing.T) {
defer metrics.DefaultRegistry.Unregister("prime.before.serving")
port := freeTCPPort(t)
l := slog.New(slog.DiscardHandler)
c := config.NewC(l)
setStatsConfig(c, map[string]any{
"type": "prometheus",
"interval": "1h",
"listen": "127.0.0.1:" + port,
"path": "/metrics",
})
gauge := metrics.GetOrRegisterGauge("prime.before.serving", nil)
// The pause would let a listener that came up before the prime serve the zero value
var s *statsServer
ctx := stopAtCleanup(t, &s)
s, err := newStatsServerFromConfig(ctx, l, c, "", false, func() {
time.Sleep(200 * time.Millisecond)
gauge.Update(42)
})
require.NoError(t, err)
go s.Start()
waitForListening(t, "127.0.0.1:"+port)
resp, err := http.Get("http://127.0.0.1:" + port + "/metrics")
require.NoError(t, err)
defer resp.Body.Close()
body, err := io.ReadAll(resp.Body)
require.NoError(t, err)
assert.Contains(t, string(body), "prime_before_serving 42")
}
// Start captures once before the first tick, the interval here is an hour
func TestStatsServer_Start_primes(t *testing.T) {
ln, err := net.Listen("tcp", "127.0.0.1:0")
require.NoError(t, err)
defer ln.Close()
l := slog.New(slog.DiscardHandler)
c := config.NewC(l)
setStatsConfig(c, map[string]any{"type": "graphite", "interval": "1h", "host": ln.Addr().String()})
primed := make(chan struct{}, 1)
var s *statsServer
ctx := stopAtCleanup(t, &s)
s, err = newStatsServerFromConfig(ctx, l, c, "", false, func() {
select {
case primed <- struct{}{}:
default:
}
})
require.NoError(t, err)
go s.Start()
select {
case <-primed:
case <-time.After(5 * time.Second):
t.Fatal("nothing was captured before the first tick")
}
}
// Graphite has to ship what emitStats set in the same pass, not the pass before
func TestStatsServer_emitStatsRunsBeforeGraphiteExport(t *testing.T) {
defer metrics.DefaultRegistry.Unregister("emitter.graphite.test")
ln, err := net.Listen("tcp", "127.0.0.1:0")
require.NoError(t, err)
defer ln.Close()
received := make(chan string, 1)
go func() {
conn, err := ln.Accept()
if err != nil {
return
}
defer conn.Close()
b, _ := io.ReadAll(conn)
received <- string(b)
}()
l := slog.New(slog.DiscardHandler)
c := config.NewC(l)
setStatsConfig(c, map[string]any{"type": "graphite", "interval": "1h", "host": ln.Addr().String()})
gauge := metrics.GetOrRegisterGauge("emitter.graphite.test", nil)
s, err := newStatsServerFromConfig(t.Context(), l, c, "", false, func() { gauge.Update(42) })
require.NoError(t, err)
s.runMu.Lock()
fns, _ := s.buildRuntime(t.Context(), *s.runCfg)
s.runMu.Unlock()
for _, fn := range fns {
fn()
}
select {
case body := <-received:
assert.Contains(t, body, ".emitter.graphite.test.value 42 ")
case <-time.After(5 * time.Second):
t.Fatal("graphite sent nothing")
}
}
// stats.interval reloads, and the runtime a reload starts still has emitStats
func TestStatsServer_reloadMovesTheInterval(t *testing.T) {
var passes atomic.Int64
l := slog.New(slog.DiscardHandler)
c := config.NewC(l)
// Nothing listens here, each export fails at once
host := "127.0.0.1:" + freeTCPPort(t)
setStatsConfig(c, map[string]any{"type": "graphite", "interval": "1h", "host": host})
var s *statsServer
ctx := stopAtCleanup(t, &s)
s, err := newStatsServerFromConfig(ctx, l, c, "", false, func() { passes.Add(1) })
require.NoError(t, err)
go s.Start()
waitFor(t, func() bool { return passes.Load() == 1 })
setStatsConfig(c, map[string]any{"type": "graphite", "interval": "20ms", "host": host})
require.NoError(t, s.reload(c, false))
waitFor(t, func() bool { return passes.Load() >= 6 })
setStatsConfig(c, map[string]any{"type": "graphite", "interval": "1h", "host": host})
require.NoError(t, s.reload(c, false))
// The reload's Stop waited out the 20ms loop. The new runtime's prime may land either side of this read, it is the
// only pass left to come
stopped := passes.Load()
time.Sleep(300 * time.Millisecond)
assert.LessOrEqual(t, passes.Load()-stopped, int64(1), "the 20ms loop kept running after the reload")
}
// A reload's new runtime never captures while the old one is still in a pass, go-metrics keeps its GC capture state
// in unguarded globals
func TestStatsServer_reloadWaitsForTheLastPass(t *testing.T) {
var active, most atomic.Int64
l := slog.New(slog.DiscardHandler)
c := config.NewC(l)
host := "127.0.0.1:" + freeTCPPort(t)
setStatsConfig(c, map[string]any{"type": "graphite", "interval": "10ms", "host": host})
var s *statsServer
ctx := stopAtCleanup(t, &s)
s, err := newStatsServerFromConfig(ctx, l, c, "", false, func() {
n := active.Add(1)
for {
m := most.Load()
if n <= m || most.CompareAndSwap(m, n) {
break
}
}
// Long enough that a reload almost always lands mid pass
time.Sleep(50 * time.Millisecond)
active.Add(-1)
})
require.NoError(t, err)
go s.Start()
for i := range 6 {
setStatsConfig(c, map[string]any{"type": "graphite", "interval": fmt.Sprintf("%dms", 11-i%2), "host": host})
require.NoError(t, s.reload(c, false))
time.Sleep(30 * time.Millisecond)
}
assert.Equal(t, int64(1), most.Load())
}
// Nothing captures once Stop returns, a pass in progress is waited out
func TestStatsServer_Stop_waitsForThePass(t *testing.T) {
var inPass atomic.Bool
entered := make(chan struct{}, 1)
l := slog.New(slog.DiscardHandler)
c := config.NewC(l)
setStatsConfig(c, map[string]any{"type": "graphite", "interval": "1h", "host": "127.0.0.1:" + freeTCPPort(t)})
var s *statsServer
ctx := stopAtCleanup(t, &s)
s, err := newStatsServerFromConfig(ctx, l, c, "", false, func() {
inPass.Store(true)
select {
case entered <- struct{}{}:
default:
}
time.Sleep(100 * time.Millisecond)
inPass.Store(false)
})
require.NoError(t, err)
go s.Start()
select {
case <-entered:
case <-time.After(5 * time.Second):
t.Fatal("never captured")
}
s.Stop()
assert.False(t, inPass.Load(), "Stop returned with a pass still running")
}
// A Start that arrives while Stop waits out the last pass doesn't capture alongside it
func TestStatsServer_Start_duringStopWaitsForThePass(t *testing.T) {
var active, most atomic.Int64
entered := make(chan struct{}, 1)
l := slog.New(slog.DiscardHandler)
c := config.NewC(l)
setStatsConfig(c, map[string]any{"type": "graphite", "interval": "1h", "host": "127.0.0.1:" + freeTCPPort(t)})
var s *statsServer
ctx := stopAtCleanup(t, &s)
s, err := newStatsServerFromConfig(ctx, l, c, "", false, func() {
n := active.Add(1)
for {
m := most.Load()
if n <= m || most.CompareAndSwap(m, n) {
break
}
}
select {
case entered <- struct{}{}:
default:
}
time.Sleep(200 * time.Millisecond)
active.Add(-1)
})
require.NoError(t, err)
go s.Start()
select {
case <-entered:
case <-time.After(5 * time.Second):
t.Fatal("never captured")
}
stopped := make(chan struct{})
go func() {
s.Stop()
close(stopped)
}()
// Stop holds runMu for its whole wait, the next Start has to arrive after it took it
waitFor(t, func() bool {
if s.runMu.TryLock() {
s.runMu.Unlock()
return false
}
return true
})
go s.Start()
select {
case <-stopped:
case <-time.After(5 * time.Second):
t.Fatal("Stop did not return")
}
waitFor(t, func() bool { return len(entered) == 1 })
assert.Equal(t, int64(1), most.Load(), "a Start captured alongside the pass Stop was waiting out")
}
// A reload that can't bind says so rather than finding out in the background, and the same config tries again once
// the port is free
func TestStatsServer_reload_bindFailureFails(t *testing.T) {
blocker, err := net.Listen("tcp", "127.0.0.1:0")
require.NoError(t, err)
port := strconv.Itoa(blocker.Addr().(*net.TCPAddr).Port)
s, c := newTestStatsServer(t)
stopAtCleanup(t, &s)
require.NoError(t, s.reload(c, true))
setStatsConfig(c, map[string]any{
"type": "prometheus",
"interval": "1s",
"listen": "127.0.0.1:" + port,
"path": "/metrics",
})
require.Error(t, s.reload(c, false))
assert.Nil(t, currentRuntime(s))
require.NoError(t, blocker.Close())
require.NoError(t, s.reload(c, false))
waitForListening(t, "127.0.0.1:"+port)
}
+29
View File
@@ -0,0 +1,29 @@
package test
import (
"flag"
"os"
"path/filepath"
"testing"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
// Update is the -update flag: Golden rewrites its file instead of comparing against it.
var Update = flag.Bool("update", false, "rewrite golden files instead of comparing against them")
// Golden compares got with the file at path, or rewrites the file when -update is set.
func Golden(t *testing.T, path, got string) {
t.Helper()
if *Update {
require.NoError(t, os.MkdirAll(filepath.Dir(path), 0755))
require.NoError(t, os.WriteFile(path, []byte(got), 0644))
return
}
want, err := os.ReadFile(path)
require.NoError(t, err, "missing golden file, run with -update")
assert.Equal(t, string(want), got)
}
+94
View File
@@ -0,0 +1,94 @@
change-remote
short: Changes the remote address used in the tunnel for the provided vpn addr
help:
flags:
-address string
The new remote address, ip:port
close-tunnel
short: Closes a tunnel for the provided vpn addr
help:
flags:
-local-only
Disables notifying the remote that the tunnel is shutting down
create-tunnel
short: Creates a tunnel for the provided vpn address
help: The lighthouses will be queried for real addresses but you can provide one as well.
flags:
-address string
Optionally provide a real remote address, ip:port
device-info
short: Prints information about the network device.
help:
flags:
-json
outputs as json with more information
-pretty
pretty prints json, assumes -json
help
short: prints available commands or help <command> for specific usage info
help:
list-hostmap
short: List all known previously connected hosts
help:
flags:
-by-index
gets all hosts in the hostmap from the index table
-json
outputs as json with more information
-pretty
pretty prints json, assumes -json
list-lighthouse-addrmap
short: List all lighthouse map entries
help:
flags:
-json
outputs as json with more information
-pretty
pretty prints json, assumes -json
list-pending-hostmap
short: List all handshaking hosts
help:
flags:
-by-index
gets all hosts in the hostmap from the index table
-json
outputs as json with more information
-pretty
pretty prints json, assumes -json
log-format
short: Gets or sets the current log format
help:
log-level
short: Gets or sets the current log level
help:
print-cert
short: Prints the current certificate being used or the certificate for the provided vpn addr
help:
flags:
-json
outputs as json
-pretty
pretty prints json, assumes -json
-raw
raw prints the PEM encoded certificate, not compatible with -json or -pretty
print-relays
short: Prints json details about all relay info
help:
flags:
-pretty
pretty prints json
print-tunnel
short: Prints json details about a tunnel for the provided vpn addr
help:
flags:
-pretty
pretty prints json
query-lighthouse
short: Query the lighthouses for the provided vpn address
help: This command is asynchronous. Only currently known udp addresses will be printed.
reload
short: Reloads configuration from disk, same as sending HUP to the process
help:
version
short: Prints the currently running version of nebula
help:
+3 -2
View File
@@ -43,7 +43,8 @@ type Conn interface {
// its own packet, so a short count means some peers were undeliverable, not that the batch failed.
// Not safe for concurrent use on the same Conn.
WriteBatch(bufs [][]byte, addrs []netip.AddrPort) (int, error)
ReloadConfig(c *config.C)
// ReloadConfig applies the listen settings that can change, r is nil at startup
ReloadConfig(c *config.C, r *config.Reload)
SupportsMultipleReaders() bool
Close() error
}
@@ -68,7 +69,7 @@ func (NoopConn) WriteTo(_ []byte, _ netip.AddrPort) error {
func (NoopConn) WriteBatch(bufs [][]byte, _ []netip.AddrPort) (int, error) {
return len(bufs), nil
}
func (NoopConn) ReloadConfig(_ *config.C) {
func (NoopConn) ReloadConfig(_ *config.C, _ *config.Reload) {
return
}
func (NoopConn) Close() error {
+6
View File
@@ -24,6 +24,12 @@ type NetworkChangeMonitor struct {
// NewNetworkChangeMonitor builds a monitor for local network changes. The returned monitor is always usable: Start
// is safe to call unconditionally, it no-ops when disabled or on a platform that does not need it.
func NewNetworkChangeMonitor(ctx context.Context, l *slog.Logger, c *config.C) *NetworkChangeMonitor {
if watchesNetworkChanges {
c.RegisterReloadCallback(func(c *config.C, r *config.Reload) {
r.RestartRequiredIfChanged("listen.rebind_on_network_change")
})
}
return &NetworkChangeMonitor{
l: l,
ctx: ctx,
Loaded 100 of 110 files, more files were not shown because too many files have changed in this diff. Show more