Seems to deflake tstest/integration tests. I can't reproduce it
anymore on one of my VMs that was consistently flaking after a dozen
runs before. Now I can run hundreds of times.
Updates #11649Fixes#7036
Change-Id: I2f7d4ae97500d507bdd78af9e92cd1242e8e44b8
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
We have seen in macOS client logs that the "operation not permitted", a
syscall.EPERM error, is being returned when traffic is attempted to be
sent. This may be caused by security software on the client.
This change will perform a rebind and restun if we receive a
syscall.EPERM error on clients running darwin. Rebinds will only be
called if we haven't performed one specifically for an EPERM error in
the past 5 seconds.
Updates #11710
Signed-off-by: Charlotte Brandhorst-Satzkorn <charlotte@tailscale.com>
Just because we don't have known endpoints for a peer does not mean that
the peer should become unreachable. If we know the peers key, it should
be able to call us, then we can talk back via whatever path it called us
on. First step - don't drop the packet in this context.
Updates tailscale/corp#19106
Signed-off-by: James Tucker <james@tailscale.com>
The netcheck package and the magicksock package coordinate via the
health package, but both sides have time based heuristics through
indirect dependencies. These were misaligned, so the implemented
heuristic aimed at reducing DERP moves while there is active traffic
were non-operational about 3/5ths of the time.
It is problematic to setup a good test for this integration presently,
so instead I added comment breadcrumbs along with the initial fix.
Updates #8603
Signed-off-by: James Tucker <james@tailscale.com>
This pretty much always results in an outage because peers won't
discover our new home region and thus won't be able to establish
connectivity.
Updates tailscale/corp#18095
Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: Ic0d09133f198b528dd40c6383b16d7663d9d37a7
Since link-local addresses are definitionally more likely to be a direct
(lower-latency, more reliable) connection than a non-link-local private
address, give those a bit of a boost when selecting endpoints.
Updates #8097
Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I93fdeb07de55ba39ba5fcee0834b579ca05c2a4e
This adds a method to wgengine.Engine and plumbed down into magicsock
to add a way to get a type-safe Tailscale-safe wrapper around a
wireguard-go device.Peer that only exposes methods that are safe for
Tailscale to use internally.
It also removes HandshakeAttempts from PeerStatusLite that was just
added as it wasn't needed yet and is now accessible ala cart as needed
from the Peer type accessor.
None of this is used yet.
Updates #7617
Change-Id: I07be0c4e6679883e6eeddf8dbed7394c9e79c5f4
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
This commit implements probing of UDP path lifetime on the tail end of
an active direct connection. Probing configuration has two parts -
Cliffs, which are various timeout cliffs of interest, and
CycleCanStartEvery, which limits how often a probing cycle can start,
per-endpoint. Initially a statically defined default configuration will
be used. The default configuration has cliffs of 10s, 30s, and 60s,
with a CycleCanStartEvery of 24h. Probing results are communicated via
clientmetric counters. Probing is off by default, and can be enabled
via control knob. Probing is purely informational and does not yet
drive any magicsock behaviors.
Updates #540
Signed-off-by: Jordan Whited <jordan@tailscale.com>
The switch in Conn.runDerpReader() on the derp.ReceivedMessage type
contained cases other than derp.ReceivedPacket that fell through to
writing to c.derpRecvCh, which should only be reached for
derp.ReceivedPacket. This can result in the last/previous
derp.ReceivedPacket to be re-handled, effectively creating a duplicate
packet. If the last derp.ReceivedPacket happens to be a
disco.CallMeMaybe it may result in a disco ping scan towards the
originating peer on the endpoints contained.
The change in this commit moves the channel write on c.derpRecvCh and
subsequent select awaiting the result into the derp.ReceivedMessage
case, preventing it from being reached from any other case. Explicit
continue statements are also added to non-derp.ReceivedPacket cases
where they were missing, in order to signal intent to the reader.
Fixes#10586
Signed-off-by: Jordan Whited <jordan@tailscale.com>
This uses the fact that we've received a frame from a given DERP region
within a certain time as a signal that the region is stil present (and
thus can still be a node's PreferredDERP / home region) even if we don't
get a STUN response from that region during a netcheck.
This should help avoid DERP flaps that occur due to losing STUN probes
while still having a valid and active TCP connection to the DERP server.
RELNOTE=Reduce home DERP flapping when there's still an active connection
Updates #8603
Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: If7da6312581e1d434d5c0811697319c621e187a0
* util/linuxfw, wgengine: allow ingress to magicsock UDP port on Linux
Updates #9084.
Currently, we have to tell users to manually open UDP ports on Linux when
certain firewalls (like ufw) are enabled. This change automates the process of
adding and updating those firewall rules as magicsock changes what port it
listens on.
Signed-off-by: Naman Sood <mail@nsood.in>
In DERP homeless mode, a DERP home connection is not sought or
maintained and the local node is not reachable.
Updates #3363
Updates tailscale/corp#396
Change-Id: Ibc30488ac2e3cfe4810733b96c2c9f10a51b8331
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
This is gated behind the silent disco control knob, which is still in
its infancy. Prior to this change disco pong reception was the only
event that could move trustBestAddrUntil forward, so even though we
weren't heartbeating, we would kick off discovery pings every
trustUDPAddrDuration and mirror to DERP.
Updates #540
Signed-off-by: Jordan Whited <jordan@tailscale.com>
This change exposes SilentDisco as a control knob, and plumbs it down to
magicsock.endpoint. No changes are being made to magicsock.endpoint
disco behavior, yet.
Updates #540
Signed-off-by: Jordan Whited <jordan@tailscale.com>
Co-authored-by: Brad Fitzpatrick <bradfitz@tailscale.com>
This change updates log messaging when cleaning up wireguard only peers.
This change also stops us unnecessarily attempting to clean up disco
pings for wireguard only endpoints.
Updates #7826
Signed-off-by: Charlotte Brandhorst-Satzkorn <charlotte@tailscale.com>
TestNewConn now passes as root on Linux. It wasn't closing the BPF
listeners and their goroutines.
The code is still a mess of two Close overlapping code paths, but that
can be refactored later. For now, make the two close paths more similar.
Updates #9945
Change-Id: I8a3cf5fb04d22ba29094243b8e645de293d9ed85
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Don't assume Linux lacks UDP_GRO support if it lacks UDP_SEGMENT
support. This mirrors a similar change in wireguard/wireguard-go@177caa7
for consistency sake. We haven't found any issues here, just being
overly paranoid.
Updates #cleanup
Signed-off-by: Jordan Whited <jordan@tailscale.com>
Regression from c15997511d. The callback could be run multiple times
from different endpoints.
Fixes#9801
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Record the number of MTU probes sent, the total bytes sent, the number of times
we got a successful return from an MTU probe of a particular size, and the max
MTU recorded.
Updates #311
Signed-off-by: Val <valerie@tailscale.com>
Automatically probe the path MTU to a peer when peer MTU is enabled, but do not
use the MTU information for anything yet.
Updates #311
Signed-off-by: Val <valerie@tailscale.com>
When sending a CLI ping with a specific size, continue to probe all possible UDP
paths to the peer until we find one with a large enough MTU to accommodate the
ping. Record any peer path MTU information we discover (but don't use it for
anything other than CLI pings).
Updates #311
Signed-off-by: Val <valerie@tailscale.com>
Add a field to record the wire MTU of the path to this address to the
addrLatency struct and rename it addrQuality.
Updates #311
Signed-off-by: Val <valerie@tailscale.com>
Prepare for path MTU discovery by splitting up the concept of
DefaultMTU() into the concepts of the Tailscale TUN MTU, MTUs of
underlying network interfaces, minimum "safe" TUN MTU, user configured
TUN MTU, probed path MTU to a peer, and maximum probed MTU. Add a set
of likely MTUs to probe.
Updates #311
Signed-off-by: Val <valerie@tailscale.com>
Prepare for path MTU discovery by splitting up the concept of
DefaultMTU() into the concepts of the Tailscale TUN MTU, MTUs of
underlying network interfaces, minimum "safe" TUN MTU, user configured
TUN MTU, probed path MTU to a peer, and maximum probed MTU. Add a set
of likely MTUs to probe.
Updates #311
Signed-off-by: Val <valerie@tailscale.com>
Replace CanPMTUD() with ShouldPMTUD() to check if peer path MTU discovery should
be enabled, in preparation for adding support for enabling/disabling peer MTU
dynamically.
Updated #311
Signed-off-by: Val <valerie@tailscale.com>
Add an enable/disable argument to setDontFragment() in preparation for dynamic
enable/disable of peer path MTU discovery. Add getDontFragment() to get the
status of the don't fragment bit from a socket.
Updates #311
Co-authored-by: James Tucker <james@tailscale.com>
Signed-off-by: Val <valerie@tailscale.com>
Use IPV6_MTU_DISCOVER for setting don't fragment on IPv6 sockets on Linux (was
using IP_MTU_DISCOVER, the IPv4 arg).
Updates #311
Signed-off-by: Val <valerie@tailscale.com>
Make the debugknob variable name for enabling peer path MTU discovery match the
env variable name.
Updates #311
Signed-off-by: Val <valerie@tailscale.com>
And convert all callers over to the methods that check SelfNode.
Now we don't have multiple ways to express things in tests (setting
fields on SelfNode vs NetworkMap, sometimes inconsistently) and don't
have multiple ways to check those two fields (often only checking one
or the other).
Updates #9443
Change-Id: I2d7ba1cf6556142d219fae2be6f484f528756e3c
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
NetworkMap.Addresses is redundant with the SelfNode.Addresses. This
works towards a TODO to delete NetworkMap.Addresses and replace it
with a method.
This is similar to #9389.
Updates #cleanup
Change-Id: Id000509ca5d16bb636401763d41bdb5f38513ba0
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
LocalBackend can talk to magicsock on its own to do this without
the "Engine" being involved.
(Continuing a little side quest of cleaning up the Engine
interface...)
Updates #cleanup
Change-Id: I8654acdca2b883b1bd557fdc0cfb90cd3a418a62
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Currently only the top four most popular changes: endpoints, DERP
home, online, and LastSeen.
Updates #1909
Change-Id: I03152da176b2b95232b56acabfb55dcdfaa16b79
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
We're trying to start using that monster type less and eventually get
rid of it.
Updates #1909
Change-Id: I8e1e725bce5324fb820a9be6c7952767863e6542
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
This is both more efficient (because the knobs' bool is only updated
whenever Node is changed, rarely) and also gets us one step closer to
removing a case of storing a netmap.NetworkMap in
magicsock. (eventually we want to phase out much of the use of that
type internally)
Updates #1909
Change-Id: I37e81789f94133175064fdc09984e4f3a431f1a1
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Previously two tsnet nodes in the same process couldn't have disjoint
sets of controlknob settings from control as both would overwrite each
other's global variables.
This plumbs a new controlknobs.Knobs type around everywhere and hangs
the knobs sent by control on that instead.
Updates #9351
Change-Id: I75338646d36813ed971b4ffad6f9a8b41ec91560
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
In prep for incremental netmap update plumbing (#1909), make peerMap
also keyed by NodeID, as all the netmap node mutations passed around
later will be keyed by NodeID.
In the process, also:
* add envknob.InDevMode, as a signal that we can panic more aggressively
in unexpected cases.
* pull two moderately large blocks of code in Conn.SetNetworkMap out
into their own methods
* convert a few more sets from maps to set.Set
Updates #1909
Change-Id: I7acdd64452ba58e9d554140ee7a8760f9043f961
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
I didn't clean up the more idiomatic map[T]bool with true values, at
least yet. I just converted the relatively awkward struct{}-valued
maps.
Updates #cleanup
Change-Id: I758abebd2bb1f64bc7a9d0f25c32298f4679c14f
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Replace %w verb with %v verb when logging errors.
Use %w only for wrapping errors with fmt.Errorf()
Fixes: #9213
Signed-off-by: Craig Rodrigues <rodrigc@crodrigues.org>
I'm not saying it works, but it compiles.
Updates #5794
Change-Id: I2f3c99732e67fe57a05edb25b758d083417f083e
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Upcoming work on incremental netmap change handling will require some
replumbing of which subsystems get notified about what. Done naively,
it could break "tailscale status --json" visibility later. To make sure
I understood the flow of all the updates I was rereading the status code
and realized parts of ipnstate.Status were being populated by the wrong
subsystems.
The engine (wireguard) and magicsock (data plane, NAT traveral) should
only populate the stuff that they uniquely know. The WireGuard bits
were fine but magicsock was populating stuff stuff that LocalBackend
could've better handled, so move it there.
Updates #1909
Change-Id: I6d1b95d19a2d1b70fbb3c875fac8ea1e169e8cb0
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
If we don't have the ICMP hint available, such as on Android, we can use
the signal of rx traffic to bias toward a particular endpoint.
We don't want to stick to a particular endpoint for a very long time
without any signals, so the sticky time is reduced to 1 second, which is
large enough to avoid excessive packet reordering in the common case,
but should be small enough that either rx provides a strong signal, or
we rotate in a user-interactive schedule to another endpoint, improving
the feel of failover to other endpoints.
Updates #8999
Co-authored-by: Charlotte Brandhorst-Satzkorn <charlotte@tailscale.com>
Signed-off-by: James Tucker <james@tailscale.com>
Signed-off-by: Charlotte Brandhorst-Satzkorn <charlotte@tailscale.com>
There are cases where we do not detect the non-viability of a route, but
we will instead observe a failure to send. In a Disco path this would
normally be handled as a side effect of Disco, which is not available to
non-Disco WireGuard nodes. In both cases, recognizing the failure as
such will result in faster convergence.
Updates #8999
Signed-off-by: James Tucker <james@tailscale.com>
LastFullPing is now used for disco or wireguard only endpoints. This
change updates the comment to make that clear.
Updates #7826
Signed-off-by: Charlotte Brandhorst-Satzkorn <charlotte@tailscale.com>
There are latency values stored in bestAddr and endpointState that are
no longer applicable after a connectivity change and should be cleared
out, following the documented behavior of the function.
Updates #8999
Signed-off-by: James Tucker <james@tailscale.com>
When sending a ping from the CLI, only accept a pong that is in reply
to the specific CLI ping we sent.
Updates #311
Signed-off-by: Val <valerie@tailscale.com>
Now a nodeAttr: ForceBackgroundSTUN, DERPRoute, TrimWGConfig,
DisableSubnetsIfPAC, DisableUPnP.
Kept support for, but also now a NodeAttr: RandomizeClientPort.
Removed: SetForceBackgroundSTUN, SetRandomizeClientPort (both never
used, sadly... never got around to them. But nodeAttrs are better
anyway), EnableSilentDisco (will be a nodeAttr later when that effort
resumes).
Updates #8923
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
To record wether user is using iptables or nftables after we add support to nftables on linux, we
are adding a field FirewallMode to NetInfo in HostInfo to reflect what firewall mode the host is
running, and form metrics. The information is gained from a global constant in hostinfo.go. We
set it when selection heuristic made the decision, and magicsock reports this to control.
Updates: tailscale/corp#13943
Signed-off-by: KevinLiang10 <kevinliang@tailscale.com>
If a node is flapping or otherwise generating lots of STUN endpoints, we
can end up caching a ton of useless values and sending them to peers.
Instead, let's apply a fixed per-Addr limit of endpoints that we cache,
so that we're only sending peers up to the N most recent.
Updates tailscale/corp#13890
Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I8079a05b44220c46da55016c0e5fc96dd2135ef8
Netcheck no longer performs I/O itself, instead it makes requests via
SendPacket and expects users to route reply traffic to
ReceiveSTUNPacket.
Netcheck gains a Standalone function that stands up sockets and
goroutines to implement I/O when used in a standalone fashion.
Magicsock now unconditionally routes STUN traffic to the netcheck.Client
that it hosts, and plumbs the send packet sink.
The CLI is updated to make use of the Standalone mode.
Fixes#8723
Signed-off-by: James Tucker <james@tailscale.com>
This sets the Don't Fragment flag, for now behind the
TS_DEBUG_ENABLE_PMTUD envknob.
Updates #311.
Signed-off-by: Val <valerie@tailscale.com>
Signed-off-by: salman <salman@tailscale.com>
This adds the capability to pad disco ping message payloads to reach a
specified size. It also plumbs it through to the tailscale ping -size
flag.
Disco pings used for actual endpoint discovery do not use this yet.
Updates #311.
Signed-off-by: salman <salman@tailscale.com>
Co-authored-by: Val <valerie@tailscale.com>
The nonce value is not read by anything, and di.sharedKey.Seal()
a few lines below generates its own. #cleanup
Signed-off-by: salman <salman@tailscale.com>
This change removes the noV4/noV6 check from addrForSendWireGuardLocked.
On Android, the client panics when reaching `rand.Intn()`, likely due to
the candidates list being containing no candidates. The suspicion is
that the `noV4` and the `noV6` are both being triggered causing the
loop to continue.
Updates tailscale/corp#12938
Updates #7826
Signed-off-by: Charlotte Brandhorst-Satzkorn <charlotte@tailscale.com>
Switch our best address selection to use a scoring-based approach, where
we boost each address based on whether it's a private IP or IPv6.
For users in cloud environments, this biases endpoint selection towards
using an endpoint that is less likely to cost the user money, and should
be less surprising to users.
This also involves updating the tests to not use private IPv4 addresses;
other than that change, the behaviour should be identical for existing
endpoints.
Updates #8097
Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I069e3b399daea28be66b81f7e44fc27b2943d8af
This change introduces address selection for wireguard only endpoints.
If a endpoint has not been used before, an address is randomly selected
to be used based on information we know about, such as if they are able
to use IPv4 or IPv6. When an address is initially selected, we also
initiate a new ICMP ping to the endpoints addresses to determine which
endpoint offers the best latency. This information is then used to
update which endpoint we should be using based on the best possible
route. If the latency is the same for a IPv4 and an IPv6 address, IPv6
will be used.
Updates #7826
Signed-off-by: Charlotte Brandhorst-Satzkorn <charlotte@tailscale.com>
Avoid selecting an endpoint as "better" than the current endpoint if the
total latency improvement is less than 1%. This adds some hysteresis to
avoid flapping between endpoints for a minimal improvement in latency.
Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: If8312e1768ea65c4b4d4e13d8de284b3825d7a73
On some platforms (notably macOS and iOS) we look up the default
interface to bind outgoing connections to. This is both duplicated
work and results in logspam when the default interface is not available
(i.e. when a phone has no connectivity, we log an error and thus cause
more things that we will try to upload and fail).
Fixed by passing around a netmon.Monitor to more places, so that we can
use its cached interface state.
Fixes#7850
Updates #7621
Signed-off-by: Mihai Parparita <mihai@tailscale.com>
We're using it in more and more places, and it's not really specific to
our use of Wireguard (and does more just link/interface monitoring).
Also removes the separate interface we had for it in sockstats -- it's
a small enough package (we already pull in all of its dependencies
via other paths) that it's not worth the extra complexity.
Updates #7621
Updates #7850
Signed-off-by: Mihai Parparita <mihai@tailscale.com>
This is a follow-up to #7905 that adds two more linters and fixes the corresponding findings. As per the previous PR, this only flags things that are "obviously" wrong, and fixes the issues found.
Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I8739bdb7bc4f75666a7385a7a26d56ec13741b7c
Previously, when updating endpoints we would immediately stop
advertising any endpoint that wasn't discovered during
determineEndpoints. This could result in, for example, a case where we
performed an incremental netcheck, didn't get any of our three STUN
packets back, and then dropped our STUN endpoint from the set of
advertised endpoints... which would result in clients falling back to a
DERP connection until the next call to determineEndpoints.
Instead, let's cache endpoints that we've discovered and continue
reporting them to clients until a timeout expires. In the above case
where we temporarily don't have a discovered STUN endpoint, we would
continue reporting the old value, then re-discover the STUN endpoint
again and continue reporting it as normal, so clients never see a
withdrawal.
Updates tailscale/coral#108
Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I42de72e7418ab328a6c732bdefc74549708cf8b9
The comment still said *magicsock.Conn implemented wireguard-go conn.Bind.
That wasn't accurate anymore.
A doc #cleanup.
Change-Id: I7fd003b939497889cc81147bfb937b93e4f6865c
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
So we're staying within the netip.Addr/AddrPort consistently and
avoiding allocs/conversions to the legacy net addr types.
Updates #5162
Change-Id: I59feba60d3de39f773e68292d759766bac98c917
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
These tests are passing locally and on CI. They had failed earlier in
the day when first fixing up CI, and it is not immediately clear why. I
have cycled IPv6 support locally, but this should not have a substantial
effect.
Updates #7876
Signed-off-by: James Tucker <jftucker@gmail.com>
