ChildPortHandshake: allow receive rights to be received

mac: ChildPortHandshake: allow receive rights to be sent

The intended use is to flip the client-server relationship in
CrashpadClient so that the initial client (parent process) furnishes the
handler process with a receive right. The parent can optionally receive
a port-destroyed notification allowing it to restart the handler if it
exits prematurely.

R=rsesek@chromium.org

Committed: https://chromium.googlesource.com/crashpad/crashpad/+/062138106c247191bb611de861ffe2cc83abc560

[Mark Mentovai, 2015-10-14 05:41:23]

Patchset #2 (id:20001) has been deleted

[Mark Mentovai, 2015-10-14 05:55:59]

Patchset #2 (id:40001) has been deleted

[Mark Mentovai, 2015-10-20 17:37:45]

mark@chromium.org changed reviewers:
+ rsesek@chromium.org

[Mark Mentovai, 2015-10-20 19:50:22]

Description was changed from

==========
mac: ChildPortHandshake: allow receive rights to be received

The intended use is to flip the client-server relationship in
CrashpadClient so that the initial client (parent process) furnishes the
handler process with a receive right. The parent can optionally receive
a port-destroyed notification allowing it to restart the handler if it
exits prematurely.
==========

to

==========
mac: ChildPortHandshake: allow receive rights to be sent

The intended use is to flip the client-server relationship in
CrashpadClient so that the initial client (parent process) furnishes the
handler process with a receive right. The parent can optionally receive
a port-destroyed notification allowing it to restart the handler if it
exits prematurely.
==========

[Robert Sesek, 2015-10-28 21:40:03]

LGTM - sorry for the delay

https://codereview.chromium.org/1408473002/diff/120001/util/mach/child_port_h...
File util/mach/child_port_handshake.cc (right):

https://codereview.chromium.org/1408473002/diff/120001/util/mach/child_port_h...
util/mach/child_port_handshake.cc:44: class ChildPortHandshakeServer : public
ChildPortServer::Interface {
final?

https://codereview.chromium.org/1408473002/diff/120001/util/mach/child_port_h...
util/mach/child_port_handshake.cc:336: PCHECK(shutdown(server_write_fd_.get(),
SHUT_RD) == 0) << "shutdown";
shutdown SHUT_RD for the error? same with SHUT_WR

https://codereview.chromium.org/1408473002/diff/120001/util/mach/child_port_h...
File util/mach/child_port_handshake.h (right):

https://codereview.chromium.org/1408473002/diff/120001/util/mach/child_port_h...
util/mach/child_port_handshake.h:82: //! or launchd interfaces.
Now that there are two ways to start both the server and client, maybe add a
short code snippet showing exemplar usage?

https://codereview.chromium.org/1408473002/diff/120001/util/mach/child_port_h...
File util/mach/child_port_handshake_test.cc (right):

https://codereview.chromium.org/1408473002/diff/120001/util/mach/child_port_h...
util/mach/child_port_handshake_test.cc:53: kClientDoesNotCheckIn,
Worth having a test for if the server dies and the client is unable to check in?

[Mark Mentovai, 2015-10-29 16:53:13]

Updated. Thanks for the review.

https://codereview.chromium.org/1408473002/diff/120001/util/mach/child_port_h...
File util/mach/child_port_handshake.h (right):

https://codereview.chromium.org/1408473002/diff/120001/util/mach/child_port_h...
util/mach/child_port_handshake.h:82: //! or launchd interfaces.
Robert Sesek wrote:
> Now that there are two ways to start both the server and client, maybe add a
> short code snippet showing exemplar usage?

A code snippet involving fork/exec and Mach is never short, but I added
something. Is this what you had in mind?

[Robert Sesek, 2015-10-29 17:45:20]

LGTM

https://codereview.chromium.org/1408473002/diff/120001/util/mach/child_port_h...
File util/mach/child_port_handshake.h (right):

https://codereview.chromium.org/1408473002/diff/120001/util/mach/child_port_h...
util/mach/child_port_handshake.h:82: //! or launchd interfaces.
On 2015/10/29 16:53:13, Mark Mentovai wrote:
> Robert Sesek wrote:
> > Now that there are two ways to start both the server and client, maybe add a
> > short code snippet showing exemplar usage?
> 
> A code snippet involving fork/exec and Mach is never short, but I added
> something. Is this what you had in mind?

Yeah, that's helpful. I guess it doesn't have to necessarily be a snippet, just
something that helps callers choose the right way to use the class.

[Mark Mentovai, 2015-10-29 18:14:21]

Committed patchset #6 (id:140001) manually as
062138106c247191bb611de861ffe2cc83abc560 (presubmit successful).