Signed-off-by: Brad Fitzpatrick <[email protected]>

Signed-off-by: Brad Fitzpatrick <[email protected]>

Signed-off-by: Jason A. Donenfeld <[email protected]>

Signed-off-by: Jason A. Donenfeld <[email protected]>

Signed-off-by: Jason A. Donenfeld <[email protected]>

Signed-off-by: Jason A. Donenfeld <[email protected]>

Signed-off-by: Jason A. Donenfeld <[email protected]>

This linked list implementation is awful, but maybe Go 2 will help
eventually, and at least we'r...

We're loosing our ownership of the port when bringing the device down,
which means another test ...

Signed-off-by: Jason A. Donenfeld <[email protected]>

In case some old packets snuck in there before, this flushes before
starting afresh.

Signed-off...

Rather than racing with Start(), since we're never destroying these
queues, we just set the vari...

Signed-off-by: Jason A. Donenfeld <[email protected]>

This lets us collect FDs even if the GC doesn't do it for us.

Signed-off-by: Jason A. Donenfeld...

Signed-off-by: Jason A. Donenfeld <[email protected]>

Signed-off-by: Jason A. Donenfeld <[email protected]>

Signed-off-by: Jason A. Donenfeld <[email protected]>

Since RoutineHandshake calls peer.SendKeepalive(), it potentially is a
writer into the encryptio...

RoutineReadFromTUN can trigger a call to SendStagedPackets.
SendStagedPackets attempts to protec...

Signed-off-by: Jason A. Donenfeld <[email protected]>

This serves two purposes.

First, it makes repeatedly stopping then starting a peer cheaper.
Sec...

Signed-off-by: Josh Bleecher Snyder <[email protected]>

The high iteration count was useful when TestUpDown
was the nexus of new bugs to investigate.

N...

On a many-core machine with the race detector enabled,
this test can take several minutes to com...

This is not necessary, and removing it speeds up detection of UAF bugs.

Signed-off-by: Josh Ble...

This matches the new naming scheme of upLocked and downLocked.

Signed-off-by: Jason A. Donenfel...

It's never used and we won't have a use for it. Also, move to go-running
stringer, for those wit...

The test previously checked the offset within a substruct, not the
offset within the allocated s...

Signed-off-by: Jason A. Donenfeld <[email protected]>

Before, the code attached a finalizer to an object that wasn't returned,
resulting in immediate ...

Without this, we wind up freeing packets that the encryption/decryption
queues still have, resul...

Signed-off-by: Jason A. Donenfeld <[email protected]>

Signed-off-by: Jason A. Donenfeld <[email protected]>

Part of being actually idempotent is that we shouldn't penalize code
that takes advantage of thi...

Signed-off-by: Jason A. Donenfeld <[email protected]>

The immediate motivation for this change is an observed deadlock.

1. A goroutine calls peer.Sto...

We have a bunch of stupid channel tricks, and I'm about to add more.
Give them their own file. T...

Signed-off-by: Josh Bleecher Snyder <[email protected]>

The old code silently accepted negative MTUs.
It also set MTUs above the maximum.
It also had ha...

We already track this state elsewhere. No need to duplicate.
The cost of calling changeState is ...

The TUN event reader does three things: Change MTU, device up, and device down.
Changing the MTU...

This commit simplifies device state management.
It creates a single unified state variable and d...

timersInit sets up the timers.
It need only be done once per peer.

timersStart does the work to...

elem.packet is always already nil.

Signed-off-by: Josh Bleecher Snyder <[email protected]>

It is no longer necessary.

Signed-off-by: Josh Bleecher Snyder <[email protected]>

It is no longer necessary, as of 454de6f3e64abd2a7bf9201579cd92eea5280996
(device: use channel c...

Signed-off-by: Jason A. Donenfeld <[email protected]>

Signed-off-by: Jason A. Donenfeld <[email protected]>

This papers over other unrelated races, unfortunately.

Signed-off-by: Jason A. Donenfeld <Jason...

The leak test had rare flakes.
If a system goroutine started at just the wrong moment, you'd get...

Signed-off-by: Jason A. Donenfeld <[email protected]>

This makes it easier to tag things onto the end manually for quick hacks.

Signed-off-by: Jason ...

Here is the old implementation:

type WaitPool struct {
c chan interface{}
}

...

Signed-off-by: Josh Bleecher Snyder <[email protected]>

Signed-off-by: Josh Bleecher Snyder <[email protected]>

Channels are FIFO which means we have guaranteed cache misses.

Signed-off-by: Jason A. Donenfel...

Signed-off-by: Jason A. Donenfeld <[email protected]>

Signed-off-by: Jason A. Donenfeld <[email protected]>

Signed-off-by: Jason A. Donenfeld <[email protected]>

Signed-off-by: Jason A. Donenfeld <[email protected]>

Suggested-by: Josh Bleecher Snyder <[email protected]>
Signed-off-by: Jason A. Donenfeld <Jason...

Signed-off-by: Jason A. Donenfeld <[email protected]>

This is what modules are for, and Go binaries can introspect.

Signed-off-by: Jason A. Donenfeld...

Signed-off-by: Jason A. Donenfeld <[email protected]>

benchmark old ns/op new ns/op delta
BenchmarkUAPIGet-16 2872 ...

It's a separate type of key that gets hashed into the chain.

Signed-off-by: Jason A. Donenfeld ...

This moves to a simple queue with no routine processing it, to reduce
scheduler pressure.

This ...

Signed-off-by: Jason A. Donenfeld <[email protected]>

Signed-off-by: Jason A. Donenfeld <[email protected]>

This makes the IpcGet method much faster.

We also refactor the traversal API to use a callback ...

There are very few cases, if any, in which a user only wants one of
these levels, so combine it ...

This commit overhauls wireguard-go's logging.

The primary, motivating change is to use a functi...

The declaration of err in

nextByte, err := buffered.ReadByte

shadows the declaration of err i...

Caught by go vet.
Signed-off-by: Josh Bleecher Snyder <[email protected]>

Plenty more to go, but a start:

name old time/op new time/op delta
UAPIGet-4 6.3...

Signed-off-by: Josh Bleecher Snyder <[email protected]>

Signed-off-by: Brad Fitzpatrick <[email protected]>

The original spec ends with \n\n especially for this reason.

Signed-off-by: Jason A. Donenfeld ...

Signed-off-by: Jason A. Donenfeld <[email protected]>

Interleaves IpcSetOperations would spell trouble.

Signed-off-by: Josh Bleecher Snyder <josh@tai...

Signed-off-by: Josh Bleecher Snyder <[email protected]>

Also, a behavior change: Stop treating a blank value as 0.
It's not in the spec.

Signed-off-by:...

Signed-off-by: Josh Bleecher Snyder <[email protected]>

Signed-off-by: Josh Bleecher Snyder <[email protected]>

Signed-off-by: Josh Bleecher Snyder <[email protected]>

The goal of this change is to make the structure
of IpcSetOperation easier to follow.

IpcSetOpe...

Unify the handling of unexpected UAPI errors.
The comment that says "should never happen" is inc...

Expand IPCError to contain a wrapped error,
and add a helper to make constructing such errors ea...

The code as written will drop any read errors on the floor.
Fix that.

Signed-off-by: Josh Bleec...

If device.NewPeer returns a nil error,
then the returned peer is always non-nil.

Signed-off-by:...

In order to avoid even the flirtation with passing on these dependencies
to ordinary consumers o...

The Go linker isn't smart enough to prevent gvisor from being pulled
into modules that use other...

Until we depend on Go 1.16 (which isn't released yet), alias our own
variable to the private mem...

Type tcpip.Error is not an error.

I've filed https://github.com/google/gvisor/issues/5314
to fi...

Signed-off-by: Josh Bleecher Snyder <[email protected]>

Signed-off-by: Josh Bleecher Snyder <[email protected]>

Minor style fix.

Signed-off-by: Josh Bleecher Snyder <[email protected]>

Newly allocated objects are already zeroed.

Signed-off-by: Josh Bleecher Snyder <josh@tailscale...

Now that we block when enqueueing to the decryption queue,
there is only one case in which we "d...

If we block when enqueuing encryption elements to the queue,
then we never drop them.

Signed-of...