Commit Graph

7852 Commits (7403d8e9a83ad69b9ab21bf9f9889e46692298d1)
 

Author SHA1 Message Date
Jonathan Nobels 02e3c046aa
net/dns: re-query system resolvers on no-upstream resolver failure on apple platforms (#12398)
Fixes tailscale/corp#20677

On macOS sleep/wake, we're encountering a condition where reconfigure the network
a little bit too quickly - before apple has set the nameservers for our interface.
This results in a persistent condition where we have no upstream resolver and
fail all forwarded DNS queries.

No upstream nameservers is a legitimate configuration, and we have no  (good) way
of determining when Apple is ready - but if we need to forward a query, and we
have no nameservers, then something has gone badly wrong and the network is
very broken.

A simple fix here is to simply inject a netMon event, which will go through the
configuration dance again when we hit the SERVFAIL condition.

Tested by artificially/randomly returning [] for the list of nameservers in the bespoke
ipn-bridge code responsible for getting the nameservers.

Signed-off-by: Jonathan Nobels <jonathan@tailscale.com>
6 months ago
Andrew Dunham d0f1a838a6 net/dnscache: use parent context to perform lookup
As an alterative to #11935 using #12003.

Updates #11935

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I05f643fe812ceeaec5f266e78e3e529cab3a1ac3
6 months ago
Mario Minardi 5f121396e9
VERSION.txt: this is v1.69.0 (#12441)
Signed-off-by: Mario Minardi <mario@tailscale.com>
6 months ago
JunYanBJSS 4c01ce9f43 tsnet: fix error formatting bug
Fixes #12411

Signed-off-by: JunYanBJSS <johnnycocoyan@hotmail.com>
6 months ago
Irbe Krumina f5936d132a
kube: fix typo (#12437)
Updates#cleanup

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
6 months ago
Irbe Krumina a95ea31a4e
kube,tailcfg: store parsed recorder tags in a separate field (#12429)
Add an additional RecorderAddrs field to tailscale.com/cap/kubernetes
capability. RecorderAddrs will only be populated by control
with the addresses of any tsrecorder tags set via Recorder.

Updates tailscale/corp#19821

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
6 months ago
Aaron Klotz 3511d1f8a2 cmd/tailscaled, net/dns, wgengine/router: start Windows child processes with DETACHED_PROCESS when I/O is being piped
When we're starting child processes on Windows that are CLI programs that
don't need to output to a console, we should pass in DETACHED_PROCESS as a
CreationFlag on SysProcAttr. This prevents the OS from even creating a console
for the child (and paying the associated time/space penalty for new conhost
processes). This is more efficient than letting the OS create the console
window and then subsequently trying to hide it, which we were doing at a few
callsites.

Fixes #12270

Signed-off-by: Aaron Klotz <aaron@tailscale.com>
6 months ago
Nick Khyl 4cdc4ed7db net/dns/resolver: return an empty successful response instead of NXDomain when resolving A records for 4via6 domains
As quad-100 is an authoritative server for 4via6 domains, it should always return responses
with a response code of 0 (indicating no error) when resolving records for these domains.
If there's no resource record of the specified type (e.g. A), it should return a response
with an empty answer section rather than NXDomain. Such a response indicates that there
is at least one RR of a different type (e.g., AAAA), suggesting the Windows stub resolver
to look for it.

Fixes tailscale/corp#20767

Signed-off-by: Nick Khyl <nickk@tailscale.com>
6 months ago
Maisem Ali 4b6a0c42c8 safesocket: add ConnectContext
This adds a variant for Connect that takes in a context.Context
which allows passing through cancellation etc by the caller.

Updates tailscale/corp#18266

Signed-off-by: Maisem Ali <maisem@tailscale.com>
6 months ago
Nick Khyl 3672f66c74 tailcfg: bump capver for NodeAttrDisableSplitDNSWhenNoCustomResolvers
Missed in b65221999c.

Updates tailscale/corp#15802

Signed-off-by: Nick Khyl <nickk@tailscale.com>
6 months ago
Andrew Dunham 93cd2ab224 util/singleflight: add DoChanContext
This is a variant of DoChan that supports context propagation, such that
the context provided to the inner function will only be canceled when
there are no more waiters for a given key. This can be used to
deduplicate expensive and cancelable calls among multiple callers
safely.

Updates #11935

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: Ibe1fb67442a854babbc6924fd8437b02cc9e7bcf
6 months ago
Irbe Krumina bc53ebd4a0
ipn/{ipnlocal,localapi},net/netkernelconf,client/tailscale,cmd/containerboot: optionally enable UDP GRO forwarding for containers (#12410)
Add a new TS_EXPERIMENTAL_ENABLE_FORWARDING_OPTIMIZATIONS env var
that can be set for tailscale/tailscale container running as
a subnet router or exit node to enable UDP GRO forwarding
for improved performance.
See https://tailscale.com/kb/1320/performance-best-practices#linux-optimizations-for-subnet-routers-and-exit-nodes
This is currently considered an experimental approach;
the configuration support is partially to allow further experimentation
with containerized environments to evaluate the performance
improvements.

Updates tailscale/tailscale#12295

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
6 months ago
Irbe Krumina 6f2bae019f
cmd/k8s-nameserver: fix AAAA record query response (#12412)
Return empty response and NOERROR for AAAA record queries
for DNS names for which we have an A record.
This is to allow for callers that might be first sending an AAAA query and then,
if that does not return a response, follow with an A record query.
Previously we were returning NOTIMPL that caused some callers
to potentially not follow with an A record query or misbehave in different ways.

Also return NXDOMAIN for AAAA record queries for names
that we DO NOT have an A record for to ensure that the callers
do not follow up with an A record query.

Returning an empty response and NOERROR is the behaviour
that RFC 4074 recommends:
https://datatracker.ietf.org/doc/html/rfc4074

Updates tailscale/tailscale#12321

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
6 months ago
Aaron Klotz df86576989 util/winutil: add AllocateContiguousBuffer and SetNTString helper funcs
AllocateContiguousBuffer is for allocating structs with trailing buffers
containing additional data. It is to be used for various Windows structures
containing pointers to data located immediately after the struct.

SetNTString performs in-place setting of windows.NTString and
windows.NTUnicodeString.

Updates #12383

Signed-off-by: Aaron Klotz <aaron@tailscale.com>
6 months ago
Irbe Krumina c3e2b7347b
tailcfg,cmd/k8s-operator,kube: move Kubernetes cap to a location that can be shared with control (#12236)
This PR is in prep of adding logic to control to be able to parse
tailscale.com/cap/kubernetes grants in control:
- moves the type definition of PeerCapabilityKubernetes cap to a location
shared with control.
- update the Kubernetes cap rule definition with fields for granting
kubectl exec session recording capabilities.
- adds a convenience function to produce tailcfg.RawMessage from an
arbitrary cap rule and a test for it.

An example grant defined via ACLs:
"grants": [{
      "src": ["tag:eng"],
      "dst": ["tag:k8s-operator"],
      "app": {
        "tailscale.com/cap/kubernetes": [{
            "recorder": ["tag:my-recorder"]
	    “enforceRecorder”: true
        }],
      },
    }
]
This grant enforces `kubectl exec` sessions from tailnet clients,
matching `tag:eng` via API server proxy matching `tag:k8s-operator`
to be recorded and recording to be sent to a tsrecorder instance,
matching `tag:my-recorder`.

The type needs to be shared with control because we want
control to parse this cap and resolve tags to peer IPs.

Updates tailscale/corp#19821

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
6 months ago
Fran Bull ba46495e11 appc: log how many routes are being written
So that we can debug customer problems more easily.

Updates #11008

Signed-off-by: Fran Bull <fran@tailscale.com>
6 months ago
Irbe Krumina 807934f00c
cmd/k8s-operator,k8s-operator: allow proxies accept advertized routes. (#12388)
Add a new .spec.tailscale.acceptRoutes field to ProxyClass,
that can be optionally set to true for the proxies to
accept routes advertized by other nodes on tailnet (equivalent of
setting --accept-routes to true).

Updates tailscale/tailscale#12322,tailscale/tailscale#10684

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
6 months ago
Irbe Krumina 53d9cac196
k8s-operator/apis/v1alpha1,cmd/k8s-operator/deploy/examples: update DNSConfig description (#11971)
Also removes hardcoded image repo/tag from example DNSConfig resource
as the operator now knows how to default those.

Updates tailscale/tailscale#11019

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
6 months ago
Tom Proctor 23e26e589f
cmd/k8s-operator,k8s-opeerator: include Connector's MagicDNS name and tailnet IPs in status (#12359)
Add new fields TailnetIPs and Hostname to Connector Status. These
contain the addresses of the Tailscale node that the operator created
for the Connector to aid debugging.

Fixes #12214

Signed-off-by: Tom Proctor <tomhjp@users.noreply.github.com>
6 months ago
Irbe Krumina 3a6d3f1a5b
cmd/k8s-operator,k8s-operator,go.{mod,sum}: make individual proxy images/image pull policies configurable (#11928)
cmd/k8s-operator,k8s-operator,go.{mod,sum}: make individual proxy images/image pull policies configurable

Allow to configure images and image pull policies for individual proxies
via ProxyClass.Spec.StatefulSet.Pod.{TailscaleContainer,TailscaleInitContainer}.Image,
and ProxyClass.Spec.StatefulSet.Pod.{TailscaleContainer,TailscaleInitContainer}.ImagePullPolicy
fields.
Document that we have images in ghcr.io on the relevant Helm chart fields.

Updates tailscale/tailscale#11675

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
6 months ago
Brad Fitzpatrick 916c4db75b net/dns: fix crash in tests
Looks like #12346 as submitted with failing tests.

Updates #12346

Change-Id: I582cd0dfb117686330d935d763d972373c5ae598
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
6 months ago
Adrian Dewhurst 0219317372 ipn/ipnlocal: improve sticky last suggestion
The last suggested exit node needs to be incorporated in the decision
making process when a new suggestion is requested, but currently it is
not quite right: it'll be used if the suggestion code has an error or a
netmap is unavailable, but it won't be used otherwise.

Instead, this makes the last suggestion into a tiebreaker when making a
random selection between equally-good options. If the last suggestion
does not make it to the final selection pool, then a different
suggestion will be made.

Since LocalBackend.SuggestExitNode is back to being a thin shim that
sets up the parameters to suggestExitNode, it no longer needs a test.
Its test was unable to be comprehensive anyway as the code being tested
contains an uncontrolled random number generator.

Updates tailscale/corp#19681

Change-Id: I94ecc9a0d1b622de3df4ef90523f1d3e67b4bfba
Signed-off-by: Adrian Dewhurst <adrian@tailscale.com>
6 months ago
Andrew Lytvynov 7a7e314096
ipn/ipnlocal,clientupdate: allow auto-updates in contaienrs (#12391)
We assume most containers are immutable and don't expect tailscale
running in them to auto-update. But there's no reason to prohibit it
outright.

Ignore the tailnet-wide default auto-update setting in containers, but
allow local users to turn on auto-updates via the CLI.

RELNOTE=Auto-updates are allowed in containers, but ignore the tailnet-wide default.

Fixes #12292

Signed-off-by: Andrew Lytvynov <awly@tailscale.com>
6 months ago
Andrea Gottardo b65221999c
tailcfg,net/dns: add controlknob to disable battery split DNS on iOS (#12346)
Updates corp#15802.

Adds the ability for control to disable the recently added change that uses split DNS in more cases on iOS. This will allow us to disable the feature if it leads to regression in production. We plan to remove this knob once we've verified that the feature works properly.

Signed-off-by: Andrea Gottardo <andrea@gottardo.me>
6 months ago
Andrew Dunham e88a5dbc92 various: fix lint warnings
Some lint warnings caught by running 'make lint' locally.

Updates #cleanup

Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I1534ed6f2f5e1eb029658906f9d62607dad98ca3
6 months ago
Aaron Klotz 34e8820301 util/winutil: add conpty package and helper for building windows.StartupInfoEx
StartupInfoBuilder is a helper for constructing StartupInfoEx structures
featuring proc/thread attribute lists. Calling its setters triggers the
appropriate setting of fields, adjusting flags as necessary, and populating
the proc/thread attribute list as necessary. Currently it supports four
features: setting std handles, setting pseudo-consoles, specifying handles
for inheritance, and specifying jobs.

The conpty package simplifies creation of pseudo-consoles, their associated
pipes, and assignment of the pty to StartupInfoEx proc/thread attributes.

Updates #12383

Signed-off-by: Aaron Klotz <aaron@tailscale.com>
6 months ago
Brad Fitzpatrick 8a11a43c28 cmd/derpprobe: support 'local' derpmap to get derp map via LocalAPI
To make it easier for people to monitor their custom DERP fleet.

Updates tailscale/corp#20654

Change-Id: Id8af22936a6d893cc7b6186d298ab794a2672524
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
6 months ago
Jordan Whited 6e106712f6
cmd/stunstamp: support probing multiple ports (#12356)
Updates tailscale/corp#20344

Signed-off-by: Jordan Whited <jordan@tailscale.com>
6 months ago
Brad Fitzpatrick 1ca323ac65 net/netcheck: fix probeProto.String result for IPv6 probes
This bug was introduced in e6b84f215 (May 2020) but was only used in
tests when stringifying probeProto values on failure so it wasn't
noticed for a long time.

But then it was moved into non-test code in 8450a18aa (Jun 2024) and I
didn't notice during the code movement that it was wrong. It's still
only used in failure paths in logs, but having wrong/ambiguous
debugging information isn't the best.

Whoops.

Updates tailscale/corp#20654

Change-Id: I296c727ed1c292a04db7b46ecc05c07fc1abc774
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
6 months ago
Brad Fitzpatrick 8450a18aa9 net/netcheck: flesh out some logging in error paths
Updates tailscale/corp#20654

Change-Id: Ie190f956b864985668f79b5b986438bbe07ce905
6 months ago
Kristoffer Dalby 95f266f1ce tsweb: add optional on completion callback func
Updates corp#17075

Co-Authored-By: Anton Tolchanov <anton@tailscale.com>
Signed-off-by: Kristoffer Dalby <kristoffer@tailscale.com>
6 months ago
Andrew Lytvynov b8cf852881
go.toolchain.rev: update to go 1.22.4 (#12365)
Updates https://github.com/tailscale/corp/issues/20635

Signed-off-by: Andrew Lytvynov <awly@tailscale.com>
6 months ago
Maisem Ali 36e8e8cd64 wgengine/magicsock: use math/rands/v2
Updates #11058

Co-authored-by: James Tucker <james@tailscale.com>
Signed-off-by: Maisem Ali <maisem@tailscale.com>
6 months ago
Fran Bull 573c8bd8c7 cmd/natc: add --wg-port flag
Updates tailscale/corp#20503

Signed-off-by: Fran Bull <fran@tailscale.com>
6 months ago
Maisem Ali 4a8cb1d9f3 all: use math/rand/v2 more
Updates #11058

Signed-off-by: Maisem Ali <maisem@tailscale.com>
6 months ago
Fran Bull d2d459d442 cmd/natc: add --ignore-destinations flag
Updates tailscale/corp#20503

Signed-off-by: Fran Bull <fran@tailscale.com>
6 months ago
Jun 9cdb33e2a4
tsnet: add a new error when HTTPS enabled but MagicDNC Disabled (#12364)
Fixes tailscale#12303

Signed-off-by: Jun <johnnycocoyan@hotmail.com>
6 months ago
Jordan Whited cf1e6c6e55
cmd/stunstamp: fix remote write retry (#12348)
Evaluation of remote write errors was using errors.Is() where it should
have been using errors.As().

Updates tailscale/corp#20344

Signed-off-by: Jordan Whited <jordan@tailscale.com>
6 months ago
Brad Fitzpatrick 6d3c10579e gokrazy: update breakglass with now-upstreamed ec2 change
This updates breakglass to use the now-upsteamed
https://github.com/gokrazy/breakglass/pull/18 change
so we're not using our fork now.

It also adds a gok wrapper tool, because doing it by hand
was tedious.

Updates #1866

Change-Id: Ifacbf5fbf0e377b3bd95c5f76c18751c2e1af7d7
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
6 months ago
Andrew Lytvynov 347e3f3d9a
go.mod,ipn/ipnlocal: update the ACME fork (#12343)
Update our fork of golang.org/x/crypto to pick up a fix for ACME ARI:
3fde5e568a

Fixes #12278

Signed-off-by: Andrew Lytvynov <awly@tailscale.com>
6 months ago
Irbe Krumina 82576190a7
tailcfg,cmd/k8s-operator: moves tailscale.com/cap/kubernetes peer cap to tailcfg (#12235)
This is done in preparation for adding kubectl
session recording rules to this capability grant that will need to
be unmarshalled by control, so will also need to be
in a shared location.

Updates tailscale/corp#19821

Signed-off-by: Irbe Krumina <irbe@tailscale.com>
6 months ago
Andrea Gottardo d636407f14
net/dns: don't set MatchDomains on Apple platforms when no upstream nameservers available (#12334)
This PR addresses a DNS issue on macOS as discussed this morning.

Signed-off-by: Andrea Gottardo <andrea@gottardo.me>
6 months ago
Adrian Dewhurst cf9f507d47 ipn/ipnlocal: only build allowed suggested node list once
Rather than building a new suggested exit node set every time, compute
it once on first use. Currently, syspolicy ensures that values do not
change without a restart anyway.

Since the set is being constructed in a separate func now, the test code
that manipulates syspolicy can live there, and the TestSuggestExitNode
can now run in parallel with other tests because it does not have global
dependencies.

Updates tailscale/corp#19681

Change-Id: Ic4bb40ccc91b671f9e542bd5ba9c96f942081515
Signed-off-by: Adrian Dewhurst <adrian@tailscale.com>
6 months ago
signed-long 1dc3136a24
cmd/k8s-operator: Support image 'repo' or 'repository' keys in helm values file (#12285)
cmd/k8s-operator/deploy/chart: Support image 'repo' or 'repository' keys in helm values

Fixes #12100

Signed-off-by: Michael Long <michaelongdev@gmail.com>
6 months ago
Andrew Lytvynov 379e2bf189
ipn/ipnlocal: stop offline auto-updates on shutdown (#12342)
Clean up the updater goroutine on shutdown, in addition to doing that on
backend state change. This fixes a goroutine leak on shutdown in tests.

Updates #cleanup
6 months ago
Jordan Whited ba0dd493c8
cmd/stunstamp: validate STUN tx ID in responses (#12339)
Extremely late arriving responses may leak across probing intervals.

Updates tailscale/corp#20344

Signed-off-by: Jordan Whited <jordan@tailscale.com>
6 months ago
Andrew Lytvynov bc4c8b65c7
ipn/ipnlocal: periodically run auto-updates when "offline" (#12118)
When the client is disconnected from control for any reason (typically
just turned off), we should still attempt to update if auto-updates are
enabled. This may help users who turn tailscale on infrequently for
accessing resources.

RELNOTE: Apply auto-updates even if the node is down or disconnected
from the coordination server.

Updates #12117

Signed-off-by: Andrew Lytvynov <awly@tailscale.com>
6 months ago
Maisem Ali 2f2f588c80 cmd/natc: use ListenPacket
Now that tsnet supports it, use it.

Updates tailscale/corp#20503

Signed-off-by: Maisem Ali <maisem@tailscale.com>
6 months ago
Brad Fitzpatrick e84751217a gokrazy: add prototype Tailscale appliance, build tooling, docs
Updates #1866

Change-Id: I546316cb833bf2919e0d6f55cdc9951f375f165b
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
6 months ago
Maisem Ali 0b1a8586eb cmd/natc: initial implementation of a NAT based connector
This adds a new prototype `cmd/natc` which can be used
to expose a services/domains to the tailnet.

It requires the user to specify a set of IPv4 prefixes
from the CGNAT range. It advertises these as normal subnet
routes. It listens for DNS on the first IP of the first range
provided to it.

When it gets a DNS query it allocates an IP for that domain
from the v4 range. Subsequent connections to the assigned IP
are then tcp proxied to the domain.

It is marked as a WIP prototype and requires the use of the
`TAILSCALE_USE_WIP_CODE` env var.

Updates tailscale/corp#20503

Signed-off-by: Maisem Ali <maisem@tailscale.com>
6 months ago