Fix Valgrind issues in dispatch_source_mach_unittest.cc.

Fix Valgrind issues in dispatch_source_mach_unittest.cc.

BUG=487674
R=thakis@chromium.org

Committed: https://crrev.com/9fb48c09a87a6607c8a27b435d33a528d39bb526
Cr-Commit-Position: refs/heads/master@{#329790}

[Robert Sesek, 2015-05-13 23:07:11]

Patchset #1 (id:1) has been deleted

[Robert Sesek, 2015-05-13 23:08:03]

Patchset #1 (id:20001) has been deleted

[Robert Sesek, 2015-05-13 23:09:39]

Issues:

InvalidRead
Invalid read of size 4
  base::DispatchSourceMachTest::port()
 
___ZN4base54DispatchSourceMachTest_NoMessagesAfterDestruction_Test8TestBodyEv_block_invoke19
  _dispatch_call_block_and_release
  _dispatch_queue_drain
  _dispatch_queue_invoke
  _dispatch_worker_thread2
  _pthread_wqthread
  start_wqthread
Address 0x1022971c0 is 16 bytes inside a block of size 24 free'd
  operator delete(void*) (m_replacemalloc/vg_replace_malloc.c:485)
 
base::DispatchSourceMachTest_NoMessagesAfterDestruction_Test::~DispatchSourceMachTest_NoMessagesAfterDestruction_Test()
  testing::Test::DeleteSelf_()


Leak_DefinitelyLost
112 bytes in 1 blocks are definitely lost in loss record 3,825 of 4,887
  calloc (m_replacemalloc/vg_replace_malloc.c:596)
  dispatch_semaphore_create
  base::DispatchSourceMachTest_ReceiveAfterResume_Test::TestBody()


Leak_DefinitelyLost
112 bytes in 1 blocks are definitely lost in loss record 3,826 of 4,887
  calloc (m_replacemalloc/vg_replace_malloc.c:596)
  dispatch_semaphore_create
  base::DispatchSourceMachTest_NoMessagesAfterDestruction_Test::TestBody()

[benwells, 2015-05-14 01:05:09]

benwells@chromium.org changed reviewers:
+ benwells@chromium.org

[benwells, 2015-05-14 01:05:10]

+cc myself (memory sheriff) and oshima (next memory sheriff).

[benwells, 2015-05-14 01:07:11]

benwells@chromium.org changed reviewers:
- benwells@chromium.org

[Nico, 2015-05-14 02:49:25]

lgtm

I guess the other test binary didn't run under valgrind? (And we don't have msan
on mac)

https://codereview.chromium.org/1132783005/diff/40001/base/mac/dispatch_sourc...
File base/mac/dispatch_source_mach_unittest.cc (right):

https://codereview.chromium.org/1132783005/diff/40001/base/mac/dispatch_sourc...
base/mac/dispatch_source_mach_unittest.cc:45: mach_port_t port = GetPort();
this is a noop, right?

[Nico, 2015-05-14 02:49:37]

The CQ bit was checked by thakis@chromium.org

[commit-bot: I haz the power, 2015-05-14 02:50:01]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1132783005/40001

[commit-bot: I haz the power, 2015-05-14 03:09:59]

Committed patchset #1 (id:40001)

[commit-bot: I haz the power, 2015-05-14 03:10:52]

Patchset 1 (id:??) landed as
https://crrev.com/9fb48c09a87a6607c8a27b435d33a528d39bb526
Cr-Commit-Position: refs/heads/master@{#329790}

[Robert Sesek, 2015-05-14 14:52:18]

On 2015/05/14 02:49:25, Nico wrote:
> lgtm
> 
> I guess the other test binary didn't run under valgrind? (And we don't have
msan
> on mac)

I guess not. Maybe our infrastructure is too hard to configure correctly? And
yes, no msan on Mac.

https://codereview.chromium.org/1132783005/diff/40001/base/mac/dispatch_sourc...
File base/mac/dispatch_source_mach_unittest.cc (right):

https://codereview.chromium.org/1132783005/diff/40001/base/mac/dispatch_sourc...
base/mac/dispatch_source_mach_unittest.cc:45: mach_port_t port = GetPort();
On 2015/05/14 02:49:25, Nico wrote:
> this is a noop, right?

For this test, yes. The change is not implicitly capturing the |this| pointer in
the block. I wanted to be consistent in the two tests, though.