Commit Graph

119 Commits

Author SHA1 Message Date
Brad Fitzpatrick 66be052a70 net/dnscache: work on IPv6-only hosts (again)
This fixes the regression where we had stopped working on IPv6-only
hosts.

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2021-01-06 19:53:13 -08:00
David Anderson 54e6c3a290 version: use OSS repo's version when building.
When building with redo, also include the git commit hash
from the proprietary repo, so that we have a precise commit
that identifies all build info (including Go toolchain version).

Add a top-level build script demonstrating to downstream distros
how to burn the right information into builds.

Adjust `tailscale version` to print commit hashes when available.

Fixes #841.

Signed-off-by: David Anderson <danderson@tailscale.com>
2020-10-28 16:17:21 -07:00
Brad Fitzpatrick 1e0be5a458 tshttp, derphttp: send Proxy-Authorization, not Authorization, to proxies
Whoops. But weirdly, sending Authorization sometimes worked?
2020-08-28 21:01:00 -07:00
Brad Fitzpatrick 28f9cd06f5 tshttpproxy, controlclient, derphttp, logpolicy: send Negotiate auth to proxies
For Windows only, and only when built with Tailscale's Go tree.

Updates tailscale/corp#583
2020-08-26 20:08:05 -07:00
Brad Fitzpatrick 169ff22a84 derp: set NotBefore and NotAfter in DERP server's metacert
Fixes regression from e415991256 that
only affected Windows users because Go only on Windows delegates x509
cert validation to the OS and Windows as unhappy with our "metacert"
lacking NotBefore and NotAfter.

Fixes #705
2020-08-24 14:57:44 -07:00
Brad Fitzpatrick e415991256 derp, derp/derphttp: remove one RTT from DERP setup
* advertise server's DERP public key following its ServerHello
* have client look for that DEPR public key in the response
  PeerCertificates
* let client advertise it's going into a "fast start" mode
  if it finds it
* modify server to support that fast start mode, just not
  sending the HTTP response header

Cuts down another round trip, bringing the latency of being able to
write our first DERP frame from SF to Bangalore from ~725ms
(3 RTT) to ~481ms (2 RTT: TCP and TLS).

Fixes #693

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2020-08-20 14:00:55 -07:00
Brad Fitzpatrick 9cb2df4ddd derp/derpmap: add London, Dallas, Seattle 2020-08-19 20:49:23 -07:00
Brad Fitzpatrick 805850add9 derp: remove JSON struct tags in comments
They don't work in comments.

Added a test too to show that there's no change in behavior.
(It does case insensitive matching on parse anyway)
2020-08-19 14:36:43 -07:00
Brad Fitzpatrick 287522730d derp/derphttp: support standard-ish SSLKEYLOGFILE environment variable
For debugging.
2020-08-18 19:23:34 -07:00
Brad Fitzpatrick c5eb57f4d6 net/tshttpproxy: new package, support WPAD/PAC proxies on Windows
Updates tailscale/corp#553

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2020-08-18 15:40:56 -07:00
Brad Fitzpatrick 93ffc565e5 derp: remove protocol version 1 support
It hasn't existed for a long time and there are no current users.

Fixes #199
2020-08-17 16:17:56 -07:00
Brad Fitzpatrick 6b80bcf112 derp: remove a client round-trip waiting on serverInfo
It just has a version number in it and it's not really needed.
Instead just return it as a normal Recv message type for those
that care (currently only tests).

Updates #150 (in that it shares the same goal: initial DERP latency)
Updates #199 (in that it removes some DERP versioning)
2020-08-17 16:15:22 -07:00
Brad Fitzpatrick e64ab89712 derp/derpmap: add Bangalore and Tokyo 2020-08-14 13:29:14 -07:00
Brad Fitzpatrick 80d0b88a89 derp/derpmap: fix constructor argument order
Fix of 3e2bfe48c3
2020-08-14 13:21:48 -07:00
Ross Zurowski 3e2bfe48c3
derpmap: add full region name
We're beginning to reference DERP region names in the admin UI, so it's
best to consolidate this information in our DERP map.

Signed-off-by: Ross Zurowski <ross@rosszurowski.com>
2020-08-14 15:57:11 -04:00
Josh Bleecher Snyder 062bd67d3b derp: use rand instead of crypto/rand to generate jitter
We don't need crypto/rand. Let the OS keep its entropy bits.

Signed-off-by: Josh Bleecher Snyder <josharian@gmail.com>
2020-08-13 14:06:50 -07:00
David Anderson 37c19970b3 derp: add a debug option to verbosely log drops to a destination.
Signed-off-by: David Anderson <danderson@tailscale.com>
2020-08-12 15:27:01 -07:00
Josh Bleecher Snyder 909c165382 derp: remove two key.Public allocations
Reading and writing a [32]byte key to a bufio.Reader/bufio.Writer
can easily by done without allocating. Do so.

It is slower; on my machine, it adds about 100ns per read/write.
However, the overall request takes a minimum of several µs,
and it cuts allocations meaningfully, so it is probably worth it.

name                      old time/op    new time/op    delta
SendRecv/msgsize=10-8       9.21µs ± 9%    9.08µs ± 8%     ~     (p=0.250 n=15+15)
SendRecv/msgsize=100-8      6.51µs ± 9%    6.60µs ± 7%     ~     (p=0.259 n=15+13)
SendRecv/msgsize=1000-8     7.24µs ±13%    7.61µs ±36%     ~     (p=1.000 n=11+15)
SendRecv/msgsize=10000-8    19.5µs ±15%    19.9µs ±25%     ~     (p=0.890 n=14+15)

name                      old speed      new speed      delta
SendRecv/msgsize=10-8     1.09MB/s ± 8%  1.10MB/s ± 8%     ~     (p=0.286 n=15+15)
SendRecv/msgsize=100-8    15.4MB/s ± 8%  15.1MB/s ± 6%     ~     (p=0.129 n=15+12)
SendRecv/msgsize=1000-8    139MB/s ±15%   135MB/s ±28%     ~     (p=1.000 n=11+15)
SendRecv/msgsize=10000-8   516MB/s ±17%   506MB/s ±21%     ~     (p=0.880 n=14+15)

name                      old alloc/op   new alloc/op   delta
SendRecv/msgsize=10-8         170B ± 1%      108B ± 1%  -36.63%  (p=0.000 n=15+15)
SendRecv/msgsize=100-8        265B ± 1%      203B ± 1%  -23.34%  (p=0.000 n=15+15)
SendRecv/msgsize=1000-8     1.18kB ± 1%    1.12kB ± 0%   -5.31%  (p=0.000 n=14+14)
SendRecv/msgsize=10000-8    18.8kB ± 2%    18.8kB ± 2%     ~     (p=0.443 n=12+12)

name                      old allocs/op  new allocs/op  delta
SendRecv/msgsize=10-8         4.00 ± 0%      2.00 ± 0%  -50.00%  (p=0.000 n=15+15)
SendRecv/msgsize=100-8        4.00 ± 0%      2.00 ± 0%  -50.00%  (p=0.000 n=15+15)
SendRecv/msgsize=1000-8       4.00 ± 0%      2.00 ± 0%  -50.00%  (p=0.000 n=15+15)
SendRecv/msgsize=10000-8      5.00 ± 0%      3.00 ± 0%  -40.00%  (p=0.000 n=13+14)

Signed-off-by: Josh Bleecher Snyder <josharian@gmail.com>
2020-08-12 15:15:58 -07:00
Brad Fitzpatrick 75225368a4 derp: fix 32-bit struct field alignment 2020-08-11 13:50:16 -07:00
David Anderson 15949ad77d derp: export the new expvar. 2020-08-11 19:59:08 +00:00
David Anderson 13661e195a derp: rename "wireguard" packet type to "other".
Strictly speaking, we don't know that it's a wireguard packet, just that
it doesn't look like a disco packet.

Signed-off-by: David Anderson <danderson@tailscale.com>
2020-08-11 19:30:39 +00:00
David Anderson 1b5b59231b derp: break down received packets by kind (disco vs. wireguard).
Signed-off-by: David Anderson <danderson@tailscale.com>
2020-08-11 19:16:28 +00:00
Josh Bleecher Snyder e577303dc7 derp: make writeUint32 and readUint32 not allocate
The allocations are small, but they're easy enough to avoid.
And it removes some clutter from the pprof output.

name                      old time/op    new time/op    delta
SendRecv/msgsize=10-8       10.1µs ± 9%     9.7µs ± 7%    -3.45%  (p=0.035 n=14+14)
SendRecv/msgsize=100-8      8.12µs ± 7%    7.38µs ± 9%    -9.02%  (p=0.000 n=15+15)
SendRecv/msgsize=1000-8     9.51µs ±25%    8.76µs ±22%      ~     (p=0.202 n=15+15)
SendRecv/msgsize=10000-8    21.1µs ±25%    19.9µs ±14%      ~     (p=0.270 n=15+14)
WriteUint32-8               25.1ns ± 4%    21.3ns ±12%   -15.01%  (p=0.000 n=14+14)
ReadUint32-8                35.4ns ± 4%    21.9ns ± 4%   -38.06%  (p=0.000 n=15+15)

name                      old alloc/op   new alloc/op   delta
SendRecv/msgsize=10-8         182B ± 2%      169B ± 1%    -7.22%  (p=0.000 n=15+13)
SendRecv/msgsize=100-8        282B ± 1%      265B ± 1%    -5.85%  (p=0.000 n=15+15)
SendRecv/msgsize=1000-8     1.19kB ± 1%    1.18kB ± 0%    -1.26%  (p=0.000 n=14+15)
SendRecv/msgsize=10000-8    19.3kB ± 4%    18.7kB ± 4%    -3.44%  (p=0.006 n=12+12)
WriteUint32-8                4.00B ± 0%     0.00B       -100.00%  (p=0.000 n=15+15)
ReadUint32-8                 4.00B ± 0%     0.00B       -100.00%  (p=0.000 n=15+15)

name                      old allocs/op  new allocs/op  delta
SendRecv/msgsize=10-8         8.00 ± 0%      4.00 ± 0%   -50.00%  (p=0.000 n=15+15)
SendRecv/msgsize=100-8        8.00 ± 0%      4.00 ± 0%   -50.00%  (p=0.000 n=15+15)
SendRecv/msgsize=1000-8       8.00 ± 0%      4.00 ± 0%   -50.00%  (p=0.000 n=15+15)
SendRecv/msgsize=10000-8      8.47 ±17%      5.00 ± 0%   -40.94%  (p=0.000 n=15+14)
WriteUint32-8                 1.00 ± 0%      0.00       -100.00%  (p=0.000 n=15+15)
ReadUint32-8                  1.00 ± 0%      0.00       -100.00%  (p=0.000 n=15+15)

Signed-off-by: Josh Bleecher Snyder <josharian@gmail.com>
2020-08-10 14:15:37 -07:00
Josh Bleecher Snyder 355c6296f0 derp: add readUint32 and writeUint32 benchmarks
These aren't particularly performance critical,
but since I have an optimization pending for them,
it's worth having a corresponding benchmark.

Signed-off-by: Josh Bleecher Snyder <josharian@gmail.com>
2020-08-10 14:15:37 -07:00
Josh Bleecher Snyder d65e2632ab derp: add basic benchmark
This benchmark is far from perfect: It mixes together
client and server. Still, it provides a starting point
for easy profiling.

Signed-off-by: Josh Bleecher Snyder <josharian@gmail.com>
2020-08-10 09:58:34 -07:00
Josh Bleecher Snyder b23f2263c1 derp: add server version to /debug, expvars
This will make it easier for a human to tell what
version is deployed, for (say) correlating line numbers
in profiles or panics to corresponding source code.

It'll also let us observe version changes in prometheus.

Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
2020-08-07 12:46:02 -07:00
David Anderson c3994fd77c derp: remove OnlyDisco option.
Active discovery lets us introspect the state of the network stack precisely
enough that it's unnecessary, and dropping the initial DERP packets greatly
slows down tests. Additionally, it's unrealistic since our production network
will never deliver _only_ discovery packets, it'll be all or nothing.

Signed-off-by: David Anderson <danderson@tailscale.com>
2020-07-27 14:09:54 -07:00
Brad Fitzpatrick 69f3ceeb7c derp/derphttp: don't return all nil from dialRegion when STUNOnly nodes 2020-07-27 10:10:10 -07:00
Brad Fitzpatrick 52969bdfb0 derp: fix atomic padding on 32-bit again
Broken by earlier OnlyDisco addition.
2020-07-16 13:38:21 -07:00
Brad Fitzpatrick a6559a8924 wgengine/magicsock: run test DERP in mode where only disco packets allowed
So we don't accidentally pass a NAT traversal test by having DERP pick up our slack
when we really just wanted DERP as an OOB messaging channel.
2020-07-16 12:58:35 -07:00
Brad Fitzpatrick c1cabe75dc derp: fix server struct fielfd alignment on 32-bit
Mostly so the GitHub CI will pass on 32-bit.
2020-07-07 09:08:15 -07:00
Brad Fitzpatrick 4732722b87 derp: add frameClosePeer to move around clients within a region
For various reasons (mostly during rollouts or config changes on our
side), nodes may end up connecting to a fallback DERP node in a
region, rather than the primary one we tell them about in the DERP
map.

Connecting to the "wrong" node is fine, but it's in our best interest
for all nodes in a domain to connect to the same node, to reduce
intra-region packet forwarding.

This adds a privileged frame type used by the control system that can
kick off a client connection when they're connected to the wrong node
in a region. Then they hopefully reconnect immediately to the correct
location. (If not, we can leave them alone and stop closing them.)

Updates tailscale/corp#372
2020-06-25 09:33:10 -07:00
Brad Fitzpatrick dd43d9bc5f derp: fix varz typo
Updates tailscale/corp#391
2020-06-25 08:43:28 -07:00
Brad Fitzpatrick b87396b5d9 cmd/derper, derp: add some more varz and consistency check handler
I'm trying to hunt down a slow drift in numbers not agreeing.
2020-06-23 14:01:51 -07:00
Brad Fitzpatrick 6fbd1abcd3 derp: update peerGone code to work with regional DERP mesh clusters too
Updates #150
Updates #388
2020-06-22 10:06:42 -07:00
Brad Fitzpatrick c8cf3169ba cmd/derper, derp/derphttp: move bulk of derp mesh code into derphttp
To be reused in various other tools.
2020-06-15 11:58:10 -07:00
Brad Fitzpatrick 6757c990a8 Fix staticcheck warning, add Makefile with staticcheck targets, lock in staticcheck version in go.mod 2020-06-15 11:05:46 -07:00
Brad Fitzpatrick abd79ea368 derp: reduce DERP memory use; don't require callers to pass in memory to use
The magicsock derpReader was holding onto 65KB for each DERP
connection forever, just in case.

Make the derp{,http}.Client be in charge of memory instead. It can
reuse its bufio.Reader buffer space.
2020-06-15 10:26:50 -07:00
Brad Fitzpatrick 07ca0c1c29 derp: fix tracking problem if conn starts local, then also joins mesh peer 2020-06-05 12:53:43 -07:00
Brad Fitzpatrick 1cb7dab881 cmd/derper: support forwarding packets amongst set of peer DERP servers
Updates #388

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2020-06-05 10:14:33 -07:00
Brad Fitzpatrick 5e0ff494a5 derp: change NewClient constructor to an option pattern
(The NewMeshClient constructor I added recently was gross in
retrospect at call sites, especially when it wasn't obvious that a
meshKey empty string meant a regular client)
2020-06-04 11:40:12 -07:00
Brad Fitzpatrick 4d599d194f derp, derp/derphttp: add key accessors, add Client.RecvDetail
Client.RecvDetail returns a connection generation so interested clients
can detect when a reconnect happened. (Will be needed for #388)
2020-06-04 11:35:53 -07:00
Brad Fitzpatrick b33c86b542 derp: add an unexported key.Public zero value variable to be less verbose 2020-06-04 11:28:00 -07:00
Brad Fitzpatrick b663ab4685 cmd/derper: treat self-connection connection watch as no-op
Updates #388
2020-06-04 08:26:05 -07:00
Brad Fitzpatrick 484b7fc9a3 derp, cmd/derper: add frameWatchConns, framePeerPresent for inter-DERP routing
This lets a trusted DERP client that knows a pre-shared key subscribe
to the connection list. Upon subscribing, they get the current set
of connected public keys, and then all changes over time.

This lets a set of DERP server peers within a region all stay connected to
each other and know which clients are connected to which nodes.

Updates #388

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2020-06-03 08:03:29 -07:00
Brad Fitzpatrick 24009241bf net/netns: move SOCKS dialing to netns for now
This lets control & logs also use SOCKS dials.

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
2020-06-01 11:00:03 -07:00
Brad Fitzpatrick cf0d19f0ab net/tlsdial, derp/derphttp: finish DERPNode.CertName validation 2020-06-01 09:01:37 -07:00
Brad Fitzpatrick becce82246 net/netns, misc tests: remove TestOnlySkipPrivilegedOps, argv checks
The netns UID check is sufficient for now. We can do something else
later if/when needed.
2020-05-31 14:40:18 -07:00
Brad Fitzpatrick 7f68e097dd net/netcheck: fix HTTPS fallback bug from earlier today
My earlier 3fa58303d0 tried to implement
the net/http.Tranhsport.DialTLSContext hook, but I didn't return a
*tls.Conn, so we ended up sending a plaintext HTTP request to an HTTPS
port. The response ended up being Go telling as such, not the
/derp/latency-check handler's response (which is currently still a
404). But we didn't even get the 404.

This happened to work well enough because Go's built-in error response
was still a valid HTTP response that we can measure for timing
purposes, but it's not a great answer. Notably, it means we wouldn't
be able to get a future handler to run server-side and count those
latency requests.
2020-05-29 22:33:08 -07:00
David Anderson 5114df415e net/netns: set the bypass socket mark on linux.
This allows tailscaled's own traffic to bypass Tailscale-managed routes,
so that things like tailscale-provided default routes don't break
tailscaled itself.

Progress on #144.

Signed-off-by: David Anderson <danderson@tailscale.com>
2020-05-29 15:16:58 -07:00