Commit Graph

7672 Commits (95f266f1ceeab1c9007ea9bc63670d3afa2d04a3)
 

Author SHA1 Message Date
Brad Fitzpatrick 614261d00d wgengine/magicsock: reset AddrSet states on Rebind
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
5 years ago
Brad Fitzpatrick c0697e1feb net/interfaces: add IsExpensive and up state to State
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
5 years ago
Brad Fitzpatrick a55c4217db wgengine: diagnose why CreateTUN might've failed for users
Fixes #273

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
5 years ago
Brad Fitzpatrick 3a3b64301e wgengine: quiet some engine reconfig logging, make more consistent
Updates #282

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
5 years ago
Dmitry Adamushko 806645ea0e derp: prevent readFrame() from reading more than len(b) bytes.
Signed-off-by: Dmitry Adamushko <da@stablebits.net>
5 years ago
Dmitry Adamushko f2c2d0de68 derp/derp_server: unregisterClient() for replaced client connections.
When unregistering a replaced client connection, move the
still-connected peers to the current client connecition. Inform
the peers that we've gone only when unregistering the active
client connection.

Signed-off-by: Dmitry Adamushko <da@stablebits.net>
5 years ago
Brad Fitzpatrick 39ae80a2e7 version: bump date 5 years ago
Brad Fitzpatrick 1ec27dbb5b wgengine: remove an allocation in reconfig
Minor.
5 years ago
Brad Fitzpatrick b24029717b ipn: outdent some code in if that's statically always true 5 years ago
David Crawshaw 5b95aa81ea wgengine: remove IpcSetOperation on LinkChange
This was only done occasionally, but was extremely disruptive
when done and is no longer necessary.

It used to be that when switching links, we had to immediately
generate handshakes to everyone we were communicating with to
punch a hole in any NAT we were talking through. (This ended up
not really working, because in the process we got rid of our
session keys and ended up having a futile conversation for many
seconds.)

Now we have DERP, our link change propogates to the other side
as a new list of endpoints, so they start spraying packets.
We will definitely get one thanks to DERP, which will cause us
to spray, opening any NAT we are behind.

The result is that for good connections, we don't trash session
keys and cause an interruption.

Signed-off-by: David Crawshaw <crawshaw@tailscale.com>
5 years ago
Brad Fitzpatrick fe149979e6 logpolicy: include Go version in version log 5 years ago
Brad Fitzpatrick 73249f8032 netcheck: aggregate spammy logging onto one concise line 5 years ago
Brad Fitzpatrick 922d9546bf wgengine: don't allocate so much in userspaceEngine.getStatus
It was one of the top garbage producers on my phone.

It's slated to be deleted and replaced anyway, but this helps in the
meantime.

The go.sum changes look scary, but the new dep only adds 240 bytes to
the binary. The go.sum noise is just cmd/go being aggressive in
including a lot of stuff (which is being fixed in Go 1.15, for what I
understand). And I ran a go mod tidy, which added some too. (I had to
write a custom wrapper around go mod tidy because this mod tidy
normally breaks on tailscale.io/control being missing but referenced
in tests)

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
5 years ago
David Crawshaw 6b2e29867e controlclient: add auth key test
Signed-off-by: David Crawshaw <crawshaw@tailscale.com>
5 years ago
David Crawshaw f8b72d2b5b cmd/tailscale: add up --authkey flag
Signed-off-by: David Crawshaw <crawshaw@tailscale.com>
5 years ago
David Crawshaw 1747d099e9 ipn: add auth key
Signed-off-by: David Crawshaw <crawshaw@tailscale.com>
5 years ago
David Crawshaw 3f087466f5 controlclient: add auth key
Signed-off-by: David Crawshaw <crawshaw@tailscale.com>
5 years ago
David Crawshaw 22a0acff39 tailcfg: add auth key to RegisterRequest
Signed-off-by: David Crawshaw <crawshaw@tailscale.com>
5 years ago
Sylvain Rabot a279032998 cmd/mkpkg: fix missing default value for --depends
Signed-off-by: Sylvain Rabot <sylvain@abstraction.fr>
5 years ago
Brad Fitzpatrick 0f64d7f9cc wgengine: remove a tiny allocation
Drop in the bucket compared to the larger issues, but noticed it
when looking at pprof.
5 years ago
Brad Fitzpatrick 59800138ff bump wireguard dep 5 years ago
Brad Fitzpatrick 277fe84c6b version: don't depend on goversion on ios 5 years ago
Brad Fitzpatrick 2776209e49 portlist: don't depend on osexec package on ios, even if it's unused
Continuation of 5bb14c07dc.

The earlier commit provided the space savings (as the linker could see
through that osexec was unused at runtime), but it didn't clean up the
dep graph (from go list -json or godepgraph).

This removes the netstat.go file from the build too, just so the dep list
looks more reasonable.
5 years ago
Brad Fitzpatrick 5bb14c07dc portlist: don't depend on tempfork/osexec on iOS, saves 90KB
This gives us 90KB more of memory on iOS, as it shrinks the
NetworkExtension binary by 90KB.

The netstat binary isn't available in the network extension anyway, so
no point pulling in the osexec package which'll just fail to find
netstat anyway.
5 years ago
Brad Fitzpatrick ec0cd60fc3 portlist: fail earlier, more nicely on iOS 5 years ago
David Anderson 96b2f20c5b types/logger: add Discard helper. 5 years ago
Dmitry Adamushko 44434fdc82 derp/derp_server: fixed unbalanced {register,unregister}Client() calls.
Signed-off-by: Dmitry Adamushko <da@stablebits.net>
5 years ago
Blake Gentry e19287f60f wgengine/magicsock: fix Conn docs type reference
The docs on magicsock.Conn stated that they implemented the
wireguard/device.Bind interface, yet this type does not exist. In
reality, the Conn type implements the wireguard/conn.Bind interface.

I also fixed a small typo in the same file.

Signed-off-by: Blake Gentry <blakesgentry@gmail.com>
5 years ago
Brad Fitzpatrick 71d6738333 tstime: change an Errorf+return to Fatalf in subtest
Forgot to git add this during review. Fail.
5 years ago
Brad Fitzpatrick febdac0499 tstime: write Parse3339 parse that doesn't use time.Parse
It doesn't allocate and it's half the time of time.Parse (which
allocates), and 2/3rds the time of time.ParseInLocation (which
doesn't).

Go with a UTC time:

BenchmarkGoParse3339/Z-8                 2200995               534 ns/op               0 B/op          0 allocs/op
BenchmarkGoParse3339/Z-8                 2254816               554 ns/op               0 B/op          0 allocs/op
BenchmarkGoParse3339/Z-8                 2159504               522 ns/op               0 B/op          0 allocs/op

Go allocates with a "-08:00" suffix instead of ending in "Z":

BenchmarkGoParse3339/TZ-8                1276491               884 ns/op             144 B/op          3 allocs/op
BenchmarkGoParse3339/TZ-8                1355858               942 ns/op             144 B/op          3 allocs/op
BenchmarkGoParse3339/TZ-8                1385484               911 ns/op             144 B/op          3 allocs/op

Go doesn't allocate if you use time.ParseInLocation, but then you need
to parse the string to find the location anyway, so might as well go
all the way (below).

BenchmarkGoParse3339InLocation-8         1912254               597 ns/op               0 B/op          0 allocs/op
BenchmarkGoParse3339InLocation-8         1980043               612 ns/op               0 B/op          0 allocs/op
BenchmarkGoParse3339InLocation-8         1891366               612 ns/op               0 B/op          0 allocs/op

Parsing RFC3339 ourselves, UTC:

BenchmarkParse3339/Z-8                   3889220               307 ns/op               0 B/op          0 allocs/op
BenchmarkParse3339/Z-8                   3718500               309 ns/op               0 B/op          0 allocs/op
BenchmarkParse3339/Z-8                   3621231               303 ns/op               0 B/op          0 allocs/op

Parsing RFC3339 ourselves, with timezone (w/ *time.Location fetched
from sync.Map)

BenchmarkParse3339/TZ-8                  3019612               418 ns/op               0 B/op          0 allocs/op
BenchmarkParse3339/TZ-8                  2921618               401 ns/op               0 B/op          0 allocs/op
BenchmarkParse3339/TZ-8                  3031671               408 ns/op               0 B/op          0 allocs/op

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
5 years ago
Brad Fitzpatrick 02948d2c86 Dockerfile: add some usage docs, bump to Go 1.14 5 years ago
Brad Fitzpatrick d503dee6f1 tstime: add new package for time utilities, starting with Parse3339
Go's time.Parse always allocates a FixedZone for time strings not in
UTC (ending in "Z"). This avoids that allocation, at the cost of
adding a cache.

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
5 years ago
David Anderson 4e0ee141e8 derp/derpmap: add derp5 (Sydney) to prod map.
Signed-off-by: David Anderson <dave@natulte.net>
5 years ago
Brad Fitzpatrick 996bf9cae7 logtail: don't send a User-Agent
Just useless bytes on the wire. Especially with HTTP/1.

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
5 years ago
Brad Fitzpatrick 7bf436ba30 tsweb: add some Benchmarks
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
5 years ago
Brad Fitzpatrick 64334143a1 tsweb: rename Handler to ReturnHandler
The name's been bugging me for a long time.

I liked neither the overlap between tsweb.Handler and http.Handler,
nor the name "ServeHTTPErr" which sounds like it's an error being
returned, like it's an error handler and not sometimes a happy path.

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
5 years ago
Brad Fitzpatrick bdc55d7091 logtail: add ParsePrivateID 5 years ago
Brad Fitzpatrick 50aeb5b9ad wgengine: unexport some windows funcs 5 years ago
David Anderson b446de103a tsweb: fix tests.
Signed-off-by: David Anderson <dave@natulte.net>
5 years ago
David Anderson af3a9dfad6 tsweb: add a StdHandler that doesn't log 200 responses.
Signed-off-by: David Anderson <dave@natulte.net>
5 years ago
David Anderson ba2774ea27 cmd/mkpkg: support depending on stuff.
Signed-off-by: David Anderson <dave@natulte.net>
5 years ago
David Anderson 362d6af4e2 cmd/tailscaled: set CacheDirectory in service.
Signed-off-by: David Anderson <dave@natulte.net>
5 years ago
Brad Fitzpatrick e71a7c7a2c logtail: read to EOF on chunked response
We'll be fixing the server so this won't trigger in practice,
but it demos the connection reuse problem.

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
5 years ago
Brad Fitzpatrick fcb6a34f4b logtail: reduce allocations encoding text
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
5 years ago
Brad Fitzpatrick 2863e49db9 tsweb: don't flush, treat no-op Handler as 200, like Go
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
5 years ago
Brad Fitzpatrick 3b4b17d239 logpolicy: log on dials, add knob to force HTTP/1 for log uploads
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
5 years ago
David Anderson 5d995d9d6b tsweb: add a test case for nil child errors in tsweb.Error.
Signed-off-by: David Anderson <dave@natulte.net>
5 years ago
David Anderson 2c2dff9559 tsweb: don't panic if we get a tsweb.Error with no embedded error.
It's technically weird to return a tsweb.Error with no child err,
but it's a sensible thing to want to do, and we shouldn't panic
if it happens.

Signed-off-by: David Anderson <dave@natulte.net>
5 years ago
David Crawshaw 2f8719741e controlclient: do not send duplicate hostinfo/netinfo
This should never happen, so log when it does so we can fix it.

Signed-off-by: David Crawshaw <crawshaw@tailscale.com>
5 years ago
David Anderson 48d7ee1c6a cmd/microproxy: adjust to export node stats and a Go expvar server's stats.
This is a temporary specialization to what tailscale prod needs right now,
it'll go back to something more generic later.
5 years ago