[mojo-edk] Fix race in port closure

[mojo-edk] Fix race in port closure

When we close a port we first collect any new ports
hanging around in its message queue. These ports have
been accepted on the node but not yet read by the
embedder, and they need to be closed to avoid leakage.

Previously we collected the ports, then released the
port/ports lock, then erased the closing port. That
left a window where a new message could be accepted
by the port on another thread and leak its ports.

This CL simply erases the closing port while still
holding locks, thus avoiding said race.

TBR=amistry@chromium.org
BUG=

Committed: https://crrev.com/bcdfd1be996817834a250fbf8595035fd48185ea
Cr-Commit-Position: refs/heads/master@{#393896}

[Ken Rockot(use gerrit already), 2016-05-16 18:27:55]

The CQ bit was checked by rockot@chromium.org

[commit-bot: I haz the power, 2016-05-16 18:28:24]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1985773002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1985773002/1

[commit-bot: I haz the power, 2016-05-16 19:23:18]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-05-16 19:25:03]

Description was changed from

==========
[mojo-edk] Fix race in port closure

When we close a port we first collect any new ports
hanging around in its message queue. These ports have
been accepted on the node but not yet read by the
embedder, and they need to be closed to avoid leakage.

Previously we collected the ports, then released the
port/ports lock, then erased the closing port. That
left a window where a new message could be accepted
by the port on another thread and leak its ports.

This CL simply erases the closing port while still
holding locks, thus avoiding said race.

TBR=amistry@chromium.org
BUG=
==========

to

==========
[mojo-edk] Fix race in port closure

When we close a port we first collect any new ports
hanging around in its message queue. These ports have
been accepted on the node but not yet read by the
embedder, and they need to be closed to avoid leakage.

Previously we collected the ports, then released the
port/ports lock, then erased the closing port. That
left a window where a new message could be accepted
by the port on another thread and leak its ports.

This CL simply erases the closing port while still
holding locks, thus avoiding said race.

TBR=amistry@chromium.org
BUG=

Committed: https://crrev.com/bcdfd1be996817834a250fbf8595035fd48185ea
Cr-Commit-Position: refs/heads/master@{#393896}
==========

[commit-bot: I haz the power, 2016-05-16 19:25:03]

Patchset 1 (id:??) landed as
https://crrev.com/bcdfd1be996817834a250fbf8595035fd48185ea
Cr-Commit-Position: refs/heads/master@{#393896}