Commit Graph

1608 Commits (01e8b7fb7ebe594825838fa74f9328697ac69644)
 

Author SHA1 Message Date
Brad Fitzpatrick 7a2a3955d3 logtail/filch: skip a broken test on Windows
Add a TODO with some notes about why it's skipped for now.

Updates #50
4 years ago
Brad Fitzpatrick a6c34bdc28 control/controlclient: also include our own profile (for when no self-owned peers)
Fix from regression in previous commit
(0e3048d8e0) that was caught by e2e
tests.

In that previous commit, the user's own profile was omitted from the
NetworkMap in the case where the user only had one node.
4 years ago
Brad Fitzpatrick 0e3048d8e0 control/controlclient: support delta userprofiles from control
I was going to make support for this advertised from the client, but
turns out only "tailscale status" even uses the UserProfiles field and
fails gracefully (omits that field) if a user profile for a user is
missing, so I think we can just reuse the DeltaPeers field from the
client to ask the control server to also delta encode the user
profiles.

For the few users running 1.1.x (unstable) versions between DeltaPeers
support (1.1.82) and this (~1.1.541), they'll just sometimes have
missing names in "tailscale status --json" or "tailscale status --web"
(the only places the UserProfile is used).
4 years ago
Brad Fitzpatrick 82f2fdc194 control/controlclient: adjust some logging point names
The previous code read too explicitly like log.Printf("I am here1"),
log.Printf("I am here2"). It still is with this change, but prettier, and
less subject to code rearranging order.
4 years ago
Brad Fitzpatrick 1fd9958e9d ipn: wait for initial portpoll result before starting controlclient
We were creating the controlclient and starting the portpoll concurrently,
which frequently resulted in the first controlclient connection being canceled
by the firsdt portpoll result ~milliseconds later, resulting in another
HTTP request.

Instead, wait a bit for the first portpoll result so it's much less likely to
interrupt our controlclient connection.

Updates tailscale/corp#557
4 years ago
Brad Fitzpatrick 1819f6f8c8 control/controlclient: set MapRequest.ReadOnly on initial empty endpoint request
On startup, clients do a MapRequest with empty endpoints while they
learn the DERP map to discover the STUN servers they then query to
learn their endpoints.

Set MapRequest.ReadOnly on those initial queries to not broadcast the
empty endpoints out to peers. The read results will come a half second
later (or less).

Updates tailscale/corp#557
4 years ago
Brad Fitzpatrick 105a820622 wgengine/magicsock: skip an endpoint update at start-up
At startup the client doesn't yet have the DERP map so can't do STUN
queries against DERP servers, so it only knows it local interface
addresses, not its STUN-mapped addresses.

We were reporting the interface-local addresses to control, getting
the DERP map, and then immediately reporting the full set of
updates. That was an extra HTTP request to control, but worse: it was
an extra broadcast from control out to all the peers in the network.

Now, skip the initial update if there are no stun results and we don't
have a DERP map.

More work remains optimizing start-up requests/map updates, but this
is a start.

Updates tailscale/corp#557
4 years ago
Brad Fitzpatrick 551e1e99e9 net/netns: don't bind to device for localhost connections
Fixes derphttp test failures on Windows (for #50).
4 years ago
Brad Fitzpatrick 746f03669c wgengine: fix lazy wireguard config bug when disco keys change
There was a bug with the lazy wireguard config code where, if the
minimum set of peers to tell wireguard didn't change, we skipped
calling userspaceEngine.updateActivityMapsLocked which updated
the various data structures that matched incoming traffic to later
reconfigure the minimum config.

That meant if an idle peer restarted and changed discovery keys, we
skipped updating our maps of disco keys/IPs that would caused us to
lazily inflate the config for that peer later if/when it did send
traffic.
4 years ago
Brad Fitzpatrick 2076a50862 wgengine/magicsock: finish a comment sentence that ended prematurely 4 years ago
Brad Fitzpatrick 371f1a9502 go.sum: add a missing entry that Go keeps adding 4 years ago
Alex Brainman f2ce64f0c6 wgengine/router: unfork winipcfg-go package, use upstream
Use golang.zx2c4.com/wireguard/windows/tunnel/winipcfg
instead of github.com/tailscale/winipcfg-go package.

Updates #760

Signed-off-by: Alex Brainman <alex.brainman@gmail.com>
4 years ago
Brad Fitzpatrick 515866d7c6 ipn, ipnserver, cmd/tailscale: add "server mode" support on Windows
This partially (but not yet fully) migrates Windows to tailscaled's
StateStore storage system.

This adds a new bool Pref, ForceDaemon, defined as:

// ForceDaemon specifies whether a platform that normally
// operates in "client mode" (that is, requires an active user
// logged in with the GUI app running) should keep running after the
// GUI ends and/or the user logs out.
//
// The only current applicable platform is Windows. This
// forced Windows to go into "server mode" where Tailscale is
// running even with no users logged in. This might also be
// used for macOS in the future. This setting has no effect
// for Linux/etc, which always operate in daemon mode.

Then, when ForceDaemon becomes true, we now write use the StateStore
to track which user started it in server mode, and store their prefs
under that key.

The ipnserver validates the connections/identities and informs that
LocalBackend which userid is currently in charge.

The GUI can then enable/disable server mode at runtime, without using
the CLI.

But the "tailscale up" CLI was also fixed, so Windows users can use
authkeys or ACL tags, etc.

Updates #275
4 years ago
Josh Bleecher Snyder d027cd81df tailcfg: restore Roles field to UserProfile 4 years ago
Brad Fitzpatrick 638127530b ipn/ipnserver: prevent use by multiple Windows users, add HTML status page
It was previously possible for two different Windows users to connect
to the IPN server at once, but it didn't really work. They mostly
stepped on each other's toes and caused chaos.

Now only one can control it, but it can be active for everybody else.

Necessary dependency step for Windows server/headless mode (#275)

While here, finish wiring up the HTTP status page on Windows, now that
all the dependent pieces are available.
4 years ago
Josh Bleecher Snyder 400e89367c tailcfg: restore Role field to MapResponse
Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Brad Fitzpatrick 22c462bd91 wgengine/monitor: fix copy/paste-o to actually monitor route changes
Due to a copy/paste-o, we were monitoring address changes twice, and
not monitoring route changes at all.

Verified with 'tailscale debug --monitor' that this actually works now (while
running 'route add 10.3.0.0 mask 255.255.0.0 10.0.0.1' and 'route delete (same)'
back and forth in cmd.exe)

In practice route changes are accompanied by address changes and this
doesn't fix any known issues. I just noticed this while reading this
code again. But at least the code does what it was trying to do now.
4 years ago
Brad Fitzpatrick 63d65368db go.mod: bump wireguard-go for x/sys/unix symbol loss
Updates golang/go#41868
4 years ago
Avery Pennarun 6332bc5e08 controlclient: print http errors if result code != 200.
Turns out for the particular error I was chasing, it actually returns
200 and zero data. But this code mirrors the same check in the map
poll, and is the right thing to do in the name of future debugging.

Signed-off-by: Avery Pennarun <apenwarr@tailscale.com>
4 years ago
Avery Pennarun 0e5f2b90a5 echoRespondToAll: filter.Accept rather than filter.Drop on a match.
This function is only called in fake mode, which won't do anything more
with the packet after we respond to it anyway, so dropping it in the
prefilter is not necessary. And it's kinda semantically wrong: we did
not reject it, so telling the upper layer that it was rejected produces
an ugly error message.

Signed-off-by: Avery Pennarun <apenwarr@tailscale.com>
4 years ago
Avery Pennarun 5041800ac6 wgengine/tstun/faketun: it's a null tunnel, not a loopback.
At some point faketun got implemented as a loopback (put a packet in
from wireguard, the same packet goes back to wireguard) which is not
useful. It's supposed to be an interface that just sinks all packets,
and then wgengine adds *only* and ICMP Echo responder as a layer on
top.

This caused extremely odd bugs on darwin, where the special case that
reinjects packets from local->local was filling the loopback channel
and creating an infinite loop (which became jammed since the reader and
writer were in the same goroutine).

Signed-off-by: Avery Pennarun <apenwarr@tailscale.com>
4 years ago
Brad Fitzpatrick 3e4c46259d wgengine/magicsock: don't do netchecks either when network is down
A continuation of 6ee219a25d

Updates #640
4 years ago
Brad Fitzpatrick 6ee219a25d ipn, wgengine, magicsock, tsdns: be quieter and less aggressive when offline
If no interfaces are up, calm down and stop spamming so much. It was
noticed as especially bad on Windows, but probably was bad
everywhere. I just have the best network conditions testing on a
Windows VM.

Updates #604
4 years ago
David Crawshaw 7616acd118 tailcfg: add Clone method for RegisterResponse
Signed-off-by: David Crawshaw <crawshaw@tailscale.com>
4 years ago
David Crawshaw 15297a3a09 control/controlclient: some extra debug info in errors
Signed-off-by: David Crawshaw <crawshaw@tailscale.com>
4 years ago
Brad Fitzpatrick 587bdc4280 ipn, wgengine: disable subnet routes if network has PAC configuration
Not configurable yet.

Updates tailscale/corp#653

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
4 years ago
Josh Bleecher Snyder a5103a4cae all: upgrade to latest version of depaware 4 years ago
Josh Bleecher Snyder 585a0d8997 all: use testing.T.TempDir
Bit of Friday cleanup.

Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Brad Fitzpatrick ed5d5f920f net/interfaces: add interfaces.State.String method 4 years ago
Josh Bleecher Snyder 9784cae23b util/uniq: add new package
This makes it easy to compact slices that contain duplicate elements
by sorting and then uniqing.

This is an alternative to constructing an intermediate map
and then extracting elements from it. It also provides
more control over equality than using a map key does.

Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Brad Fitzpatrick 12e28aa87d ipn: on transition from no PAC to PAC, reset state
So previous routes aren't shadowing resources that the operating
system might need (Windows Domain Controller, DNS server, corp HTTP
proxy, WinHTTP fetching the PAC file itself, etc).

This effectively detects when we're transitioning from, say, public
wifi to corp wifi and makes Tailscale remove all its routes and stops
its TCP connections and tries connecting to everything anew.

Updates tailscale/corp#653
4 years ago
Brad Fitzpatrick cab3eb995f net/interfaces: quiet PAC detection logging in no-PAC case, add benchmark 4 years ago
Josh Bleecher Snyder 38dda1ea9e all: update depaware.txt
Broken by 8051ecff55.


Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Brad Fitzpatrick 8051ecff55 net/interfaces: add State.PAC field, populate it on Windows
Not used for anything yet (except logging), but populate the current
proxy autoconfig PAC URL in Interfaces.State.

A future change will do things based on it.
4 years ago
Brad Fitzpatrick b5a3850d29 control/controlclient, ipn: store machine key separately from user prefs/persist
Updates #610 (fixes after some win/xcode changes in a separate repo)
4 years ago
Josh Bleecher Snyder e1596d655a tstest: skip resource check when test has failed
Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Josh Bleecher Snyder ce6aca13f0 tailcfg: add yet another IsZero method
Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Josh Bleecher Snyder 070dfa0c3d tailcfg: add more IsZero methods
Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Josh Bleecher Snyder efb08e4fee all: use IsZero methods
Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Brad Fitzpatrick c8f257df00 Revert "all: keep UserProfiles a slice instead of a map for longer"
This reverts commit e5894aba42.

Breaks macOS/iOS build. Reverting per chat with Josh; he'll fix later today.
4 years ago
Brad Fitzpatrick 90b7293b3b ipn: add/move some constants, update a comment
And make the StateStore implementations be Stringers, for error messages.
4 years ago
Josh Bleecher Snyder 1fecf87363 control/controlclient: use wgcfg.PrivateKey.IsZero
Generated by eg using template:

---

package p

import "github.com/tailscale/wireguard-go/wgcfg"

func before(k wgcfg.PrivateKey) bool { return k == wgcfg.PrivateKey{} }
func after(k wgcfg.PrivateKey) bool  { return k.IsZero() }


Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Josh Bleecher Snyder 2b8d2babfa tailcfg: add IsZero methods to UserID and NodeID
These will be helpful for doing some automated refactoring.

Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Josh Bleecher Snyder e5894aba42 all: keep UserProfiles a slice instead of a map for longer
Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Josh Bleecher Snyder 4d4ca2e496 control/controlclient: remove Roles fields from client
They are unused.

Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
David Anderson c493e5804f wgengine/router: make v6-ness configurable in test, for consistent results.
Signed-off-by: David Anderson <danderson@tailscale.com>
4 years ago
Josh Bleecher Snyder d3701417fc tailcfg: fix typo in comment
Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Brad Fitzpatrick c86761cfd1 Remove tuntap references. We only use TUN. 4 years ago
Brad Fitzpatrick 8b94a769be cmd/tailscaled: use the standard flag page instead of getopt
Per discussion with @crawshaw. The CLI tool already used std flag anyway.
If either of them, it would've made more sense for the CLI to use getopt.
4 years ago
Brad Fitzpatrick 94a68a113b go.sum: tidy 4 years ago