Commit Graph

1330 Commits (a3f17b81083db5bb2ef90ef2a1e0a2270fe101ee)
 

Author SHA1 Message Date
Christina Wen f0e9dcdc0a
wgengine/router: restore /etc/resolv.conf after tailscale down is called
This change is to restore /etc/resolv.conf after tailscale down is called. This is done by setting the dns.Manager before errors occur. Error collection is also added.

Fixes #723
4 years ago
Brad Fitzpatrick 904a91038a tailcfg: add MapRequest.ReadOnly and OmitPeers; remove DebugForceDisco
DebugForceDisco was a development & safety knob during the the transition
to discovery. It's no longer needed.

Add MapRequest.ReadOnly to prevent clients needing to do two
peer-spamming MapRequest at start-up.

This only adds the field, not the use of the field. (The control server
needs to support it first.)

Updates tailscale/corp#557

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
4 years ago
Brad Fitzpatrick c41947903a ipn: don't log if legacy prefs don't exist (the normal case these days) 4 years ago
David Crawshaw 815bf017fc tsweb: when unwrapping HTTPError, record the user-facing message also in the log
There's often some useful piece of information in there not already
repeated in the internal error.

Signed-off-by: David Crawshaw <crawshaw@tailscale.com>
4 years ago
David Crawshaw dea3ef0597 tsweb: make JSONHandlerFunc implement ReturnHandler, not http.Handler
This way something is capable of logging errors on the server.

Fixes #766

Signed-off-by: David Crawshaw <crawshaw@tailscale.com>
4 years ago
Brad Fitzpatrick 3aeb2e204c go.mod: bump github.com/tailscale/winipcfg-go for some Windows fixes 4 years ago
Brad Fitzpatrick acafe9811f wgengine/router: run netsh advfirewall less, rename, document setFirewall 4 years ago
Christina Wen 48fbe93e72
wgengine/magicsock: clarify pre-disco 'tailscale ping' error message
This change clarifies the error message when a user pings a peer that is using an outdated version of Tailscale.
4 years ago
Brad Fitzpatrick 96fd20e3c0 ipn: bail out a bit earlier when a peer doesn't have a DNS name
It's properly handled later in tsdns.NewMap anyway, but there's work
done in the meantime that can be skipped when a peer lacks a DNS name.
It's also more clear that it's okay for it to be blank.
4 years ago
Josh Bleecher Snyder 7f97cf654d cmd/microproxy: add -insecure flag
This makes it easier to run microproxy locally during development.

Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Josh Bleecher Snyder 3fa863e6d9 cmd/derper: add missing html.EscapeString calls in /debug page
Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Christina Wen e862f90e34
wgengine/router/router_linux.go: fixed " route del failed" error (#756)
* wgengine/router/router_linux.go: Switched `cidrDiff("addr")` and `cidrDiff("route")` order

Signed-off-by: Christina Wen <christina@tailscale.com>

Co-authored-by: Christina Wen <christina@tailscale.com>
4 years ago
Brad Fitzpatrick 761fe19e5f control/controlclient: don't accept /32 routes without --accept-routes
Fixes tailscale/corp#500

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
4 years ago
Brad Fitzpatrick 88107b1287 control/controlclient: add TS_DEBUG_MAPRESPONSE to dump all MapResponses
I'm always adding this by hand during development. Time to check it in.
4 years ago
Brad Fitzpatrick 931bcd44cb control/controlclient: report Synology "distro" + its version to control 4 years ago
Brad Fitzpatrick 7e9d1f7808 wgengine/router: tighten isMissingIPv6Err
So it doesn't false positive if misused.

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
4 years ago
David Anderson 8f5b52e571 net/netns: add windows support.
Also remove rebinding logic from the windows router. Magicsock will
instead rebind based on link change signals.

Signed-off-by: David Anderson <danderson@tailscale.com>
4 years ago
Josh Bleecher Snyder 3f4d93feb2 go.mod: bump depaware to get diffs out of -check, again
I had to use

go get -u github.com/tailscale/depaware@e09ee10c18249e4bf198e66bbd47babcd502637a

to force it to the correct version; it kept selecting head~1.

Maybe because the branch is called main instead of master?
Maybe because of some delay?
4 years ago
Josh Bleecher Snyder a5d701095b wgengine/magicsock: increase test timeout to reduce flakiness
Updates #654. See that issue for a discussion of why
this timeout reduces flakiness, and what next steps are.

Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Josh Bleecher Snyder 0c0239242c wgengine/magicsock: make discoPingPurpose a stringer
It was useful for debugging once, it'll probably be useful again.

Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Josh Bleecher Snyder 6e38d29485 wgengine/magicsock: improve test logging output
This fixes line numbers and reduces timestamp
precision to overwhelming the output.

Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Josh Bleecher Snyder 41f6c78c53 go.mod: bump depaware to get diffs out of -check 4 years ago
Brad Fitzpatrick 662c19551a control/controlclient: deal with localized 'Version' string getting Windows version 4 years ago
Brad Fitzpatrick 4f7751e025 Update depaware for previous ipnserver change. 4 years ago
Brad Fitzpatrick 4f71319f7c ipn/ipnserver: make ipnserver also be an HTTP server for localhost clients
For now it just says hello to show auth works. More later.
4 years ago
Brad Fitzpatrick 3af64765fd ipn: fix Windows crash from improperly strict assertion 4 years ago
Brad Fitzpatrick a084c44afc wgengine, wgengine/router, cmd/tailscale: force netfilter mode off on Synology
For now. Get it working again so it's not stuck on 0.98.

Subnet relay can come later.

Updates #451

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
4 years ago
Brad Fitzpatrick 31c13013ae wgengine/router: tolerate disabled IPv6 on Windows
Fixes #412
Updates #524 (maybe fixes?)

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
4 years ago
Josh Bleecher Snyder 9ab2b32569 syncs: add Watch, for monitoring mutex contention
Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Brad Fitzpatrick 5a94317628 Makefile: remove tsshd from depaware
I'd removed it from the GitHub actions checks before submitted earlier
but forgot to update it here.
4 years ago
Josh Bleecher Snyder 37b40b035b wgengine/router/dns: appease staticcheck (again) 4 years ago
Brad Fitzpatrick bc1751a376 util/pidowner: add missing copyright header 4 years ago
Brad Fitzpatrick b14288f96c util/pidowner: add two missing copyright headers 4 years ago
Brad Fitzpatrick 23f01174ea util/pidowner: new package to map from process ID to its user ID 4 years ago
Brad Fitzpatrick 40e12c17ec net/netcheck: remove mistaken double Mutex.Unlock in error path
Thanks to @dotaheor for noticing.

Fixes #751
4 years ago
Brad Fitzpatrick f65eb4e5c1 net/netstat: start of new netstat package, with Windows for now
This will be used in a future change to do localhost connection
authentication. This lets us quickly map a localhost TCP connection to
a PID. (A future change will then map a pid to a user)

TODO: pull portlist's netstat code into this package. Then portlist
will be fast on Windows without requiring shelling out to netstat.exe.
4 years ago
Brad Fitzpatrick 8b60936913 depaware: update deps 4 years ago
Brad Fitzpatrick edb47b98a8 scripts: don't descend into .git for license header check 4 years ago
Josh Bleecher Snyder a877dd575c wgengine/router/dns: remove unnecessary lint ignore line
Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Josh Bleecher Snyder bf24d54143 syncs: add AssertLocked
This allows us to check lock invariants.

It was proposed upstream and rejected in:
https://github.com/golang/go/issues/1366

Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Brad Fitzpatrick 158202dbb1 go mod tidy 4 years ago
Brad Fitzpatrick 7795fcf464 Add tooldeps package to keep depaware pinned in go.mod. 4 years ago
Brad Fitzpatrick 22ed3c503e
Add depaware.txt files and GitHub checks. (#745)
See https://github.com/tailscale/depaware

Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
4 years ago
Josh Bleecher Snyder 2e40c4b564 tstest: don't log on success
This is particularly important for benchmarks,
where stray output disrupts tooling.

Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Brad Fitzpatrick 913c1bd04f wgengine/router: on Windows, add a firewall rule to allow Tailscale
Updates tailscale/tailscale#454
4 years ago
Emmanuel T Odeke 688f923db1
log/logheap: properly document LogHeap as performing HTTP upload (#741)
LogHeap no longer logs to os.Stderr and instead uploads
the heap profile by means of an HTTP POST request to the
target URL endpoint.

While here, also ensured that an error from pprof.WriteHeapProfile
isn't ignored and will prevent the HTTP request from being made
if non-nil.

Signed-off-by: Emmanuel T Odeke <emmanuel@orijtech.com>
4 years ago
Josh Bleecher Snyder 96160973ce tailcfg: regenerate Clone methods
cmd/cloner has changed. Regenerate.

Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Josh Bleecher Snyder 7bd89359c9 cmd/cloner: generate a package-level Clone function
This Clone function knows how to clone any types
for which it has generated Clone methods.
This allows callers to efficiently clone
an inbound interface{} that might contain one of these types.

Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Josh Bleecher Snyder 99d223130c cmd/cloner: fix found type detection
This was causing any type to be reported as found,
as long as there were any type decls at all. Oops.

Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago
Josh Bleecher Snyder 2352690bde cmd/cloner: ensure cloner gets re-run when structs change
If you change a struct and don't re-run cloner,
your Cloner method might be inaccurate, leading to bad things.

To prevent this, write out the struct as it is at the moment that
cloner is caller, and attempt a conversion from that type.
If the struct gets changed in any way, this conversion will fail.

This will yield false positives: If you change a non-pointer field,
you will be forced to re-run cloner, even though the actual generated
code won't change. I think this is an acceptable cost: It is a minor
annoyance, which will prevent real bugs.

Signed-off-by: Josh Bleecher Snyder <josh@tailscale.com>
4 years ago