This only adds the field, to be used in a future commit.
Updates tailscale/corp#8020
Co-authored-by: Melanie Warrick <warrick@tailscale.com>
Signed-off-by: Maisem Ali <maisem@tailscale.com>
Previously it would dial out using the http.DefaultClient, however that doesn't work
when tailscaled is running in userspace mode (e.g. when testing).
Updates tailscale/corp#9967
Signed-off-by: Maisem Ali <maisem@tailscale.com>
Kubernetes uses SPDY/3.1 which is incompatible with HTTP/2, disable it
in the transport and server.
Fixes#7645Fixes#7646
Signed-off-by: Maisem Ali <maisem@tailscale.com>
Kubernetes doesn't allow slashes as keys in secrets, replace them with "__".
This shows up in the kubernetes-operator now that tsnet sets resets the ServeConfig
at startup.
Fixes#7662
Signed-off-by: Maisem Ali <maisem@tailscale.com>
This change focuses on the backend log ID, which is the mostly commonly
used in the client. Tests which don't seem to make use of the log ID
just use the zero value.
Signed-off-by: Will Norris <will@tailscale.com>
Move the assertions about our post-privilege-drop UID/GID out of the
conditional if statement and always run them; I haven't been able to
find a case where this would fail. Defensively add an envknob to disable
this feature, however, which we can remove after the 1.40 release.
Updates #7616
Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: Iaec3dba9248131920204bd6c6d34bbc57a148185
This makes it less likely that we trip over bugs like golang/go#1435.
Updates #7616
Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: Ic28c03c3ad8ed5274a795c766b767fa876029f0e
Otherwise we see errors like
```
ssh-session(sess-20230322T005655-5562985593): recording: error sending recording to <addr>:80: Post "http://<addr>:80/record": context canceled
```
The ss.ctx is closed when the session closes, but we don't want to break the upload at that time. Instead we want to wait for the session to
close the writer when it finishes, which it is already doing.
Updates tailscale/corp#9967
Signed-off-by: Maisem Ali <maisem@tailscale.com>
Switch to using logtail for logging sockstat logs. Always log locally
(on supported platforms), but disable automatic uploading. Change
existing c2n sockstats request to trigger upload to log server and
return log ID.
Signed-off-by: Will Norris <will@tailscale.com>
Allows the iOS and macOS apps to include their frontend logs when
generating bug reports (tailscale/corp#9982).
Signed-off-by: Mihai Parparita <mihai@tailscale.com>
Currently we only send down recorders in first action, allow the final action
to replace them but not to drop them.
Updates tailscale/corp#9967
Signed-off-by: Maisem Ali <maisem@tailscale.com>
This change introduces the Recorders field to the SSHRule struct. The
field is used to store and define addresses where the ssh recorder is
located.
Signed-off-by: Charlotte Brandhorst-Satzkorn <charlotte@tailscale.com>
Previously, the build ended up embedding an empty string, which made
the shell wrapper rebuild gocross on every invocation. This is still
reasonably fast, but fixing the bypass shaves 80% off gocross's overhead
when no rebuild is needed.
Signed-off-by: David Anderson <danderson@tailscale.com>
Followup to #7518 to also export client metrics when the active interface
is cellular.
Updates tailscale/corp#9230
Updates #3363
Signed-off-by: Mihai Parparita <mihai@tailscale.com>
A bunch of us invoke tool/go from outside the repo that hosts gocross,
as a way of accessing our version-controlled toolchain. This removes
assumptions from gocross that it's being invoked within the repository
that contains its source code and toolchain configuration.
Fixestailscale/corp#9627
Signed-off-by: David Anderson <danderson@tailscale.com>
This used to make sense, but after a refactor somewhere along the line
this results in trying to download from a malformed URL and generally
confusing failures.
Signed-off-by: David Anderson <danderson@tailscale.com>
In May 2021, Azure App Services used 172.16.x.x addresses:
```
10: eth0@if11: <BROADCAST,MULTICAST,UP,LOWER_UP,M-DOWN> mtu 1500 qdisc noqueue state UP
link/ether 02:42:ac:10:01:03 brd ff:ff:ff:ff:ff:ff
inet 172.16.1.3/24 brd 172.16.1.255 scope global eth0
valid_lft forever preferred_lft forever
```
Now it uses link-local:
```
2: eth0@if6: <BROADCAST,MULTICAST,UP,LOWER_UP,M-DOWN> mtu 1500 qdisc noqueue state UP
link/ether 8a:30:1f:50:1d:23 brd ff:ff:ff:ff:ff:ff
inet 169.254.129.3/24 brd 169.254.129.255 scope global eth0
valid_lft forever preferred_lft forever
```
This is reasonable for them to choose to do, it just broke the handling in net/interfaces.
This PR proposes to:
1. Always allow link-local in LocalAddresses() if we have no better
address available.
2. Continue to make isUsableV4() conditional on an environment we know
requires it.
I don't love the idea of having to discover these environments one by
one, but I don't understand the consequences of making isUsableV4()
return true unconditionally. It makes isUsableV4() essentially always
return true and perform no function.
Fixes https://github.com/tailscale/tailscale/issues/7603
Signed-off-by: Denton Gentry <dgentry@tailscale.com>
On FreeBSD and Darwin, changing a process's supplementary groups with
setgroups(2) will also change the egid of the process, setting it to the
first entry in the provided list. This is distinct from the behaviour on
other platforms (and possibly a violation of the POSIX standard).
Because of this, on FreeBSD with no TTY, our incubator code would
previously not change the process's gid, because it would read the
newly-changed egid, compare it against the expected egid, and since they
matched, not change the gid. Because we didn't use the 'login' program
on FreeBSD without a TTY, this would propagate to a child process.
This could be observed by running "id -p" in two contexts. The expected
output, and the output returned when running from a SSH shell, is:
andrew@freebsd:~ $ id -p
uid andrew
groups andrew
However, when run via "ssh andrew@freebsd id -p", the output would be:
$ ssh andrew@freebsd id -p
login root
uid andrew
rgid wheel
groups andrew
(this could also be observed via "id -g -r" to print just the gid)
We fix this by pulling the details of privilege dropping out into their
own function and prepending the expected gid to the start of the list on
Darwin and FreeBSD.
Finally, we add some tests that run a child process, drop privileges,
and assert that the final UID/GID/additional groups are what we expect.
More information can be found in the following article:
https://www.usenix.org/system/files/login/articles/325-tsafrir.pdf
Updates #7616
Alternative to #7609
Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I0e6513c31b121108b50fe561c89e5816d84a45b9
This allows tracking packet flow via logs for prober clients. Note that
the new sclient.debug() function is called on every received packet, but
will do nothing for most clients.
I have adjusted sclient logging to print public keys in short format
rather than full. This takes effect even for existing non-debug logging
(mostly client disconnect messages).
Example logs for a packet being sent from client [SbsJn] (connected to
derper [dM2E3]) to client [10WOo] (connected to derper [AVxvv]):
```
derper [dM2E3]:
derp client 10.0.0.1:35470[SbsJn]: register single client mesh("10.0.1.1"): 4 peers
derp client 10.0.0.1:35470[SbsJn]: read frame type 4 len 40 err <nil>
derp client 10.0.0.1:35470[SbsJn]: SendPacket for [10WOo], forwarding via <derphttp_client.Client [AVxvv] url=https://10.0.1.1/derp>: <nil>
derp client 10.0.0.1:35470[SbsJn]: read frame type 0 len 0 err EOF
derp client 10.0.0.1:35470[SbsJn]: read EOF
derp client 10.0.0.1:35470[SbsJn]: sender failed: context canceled
derp client 10.0.0.1:35470[SbsJn]: removing connection
derper [AVxvv]:
derp client 10.0.1.1:50650[10WOo]: register single client
derp client 10.0.1.1:50650[10WOo]: received forwarded packet from [SbsJn] via [dM2E3]
derp client 10.0.1.1:50650[10WOo]: sendPkt attempt 0 enqueued
derp client 10.0.1.1:50650[10WOo]: sendPacket from [SbsJn]: <nil>
derp client 10.0.1.1:50650[10WOo]: read frame type 0 len 0 err EOF
derp client 10.0.1.1:50650[10WOo]: read EOF
derp client 10.0.1.1:50650[10WOo]: sender failed: context canceled
derp client 10.0.1.1:50650[10WOo]: removing connection
```
Signed-off-by: Anton Tolchanov <anton@tailscale.com>
This allows disabling spread mode, which is helpful if you are manually
running derpprobe in `--once` mode against a small number of DERP
machines.
Updates https://github.com/tailscale/corp/issues/9916
Signed-off-by: Anton Tolchanov <anton@tailscale.com>
Xcode changed how/what data it exports to build steps at some point
recently, so our old way of figuring out the minimum support version
for clang stopped working.
Updates tailscale/corp#4095
Signed-off-by: David Anderson <danderson@tailscale.com>
Sometimes, our cached toolchain ends up being an older version of
Go, older than our go.mod allows. In that scenario, gocross-wrapper.sh
would find a usable toolchain, but then fail to compile gocross.
This change makes the wrapper script check that the cached toolchain's
minor version is good enough to build tailscale.com, and re-bootstraps
in shell if not.
Signed-off-by: David Anderson <danderson@tailscale.com>
They're not needed for the sockstats logger, and they're somewhat
expensive to return (since they involve the creation of a map per
label). We now have a separate GetInterfaces() method that returns
them instead (which we can still use in the PeerAPI debug endpoint).
If changing sockstatlog to sample at 10,000 Hz (instead of the default
of 10Hz), the CPU usage would go up to 59% on a iPhone XS. Removing the
per-interface stats drops it to 20% (a no-op implementation of Get that
returns a fixed value is 16%).
Updates tailscale/corp#9230
Updates #3363
Signed-off-by: Mihai Parparita <mihai@tailscale.com>
I thought our versioning scheme would make go.mod include a commit hash
even on stable builds. I was wrong. Fortunately, the rest of this code
wants anything that 'git rev-parse' understands (to convert it into a full
git hash), and tags qualify.
Signed-off-by: David Anderson <danderson@tailscale.com>
tsnet.Server.Close was calling listener.Close with the server mutex
held, but the listener close method tries to grab that mutex, resulting
in a deadlock.
Co-authored-by: David Crawshaw <crawshaw@tailscale.com>
Signed-off-by: Maisem Ali <maisem@tailscale.com>
