Properly create zero-initialized local memory.

Properly create zero-initialized local memory.

Previous code used memset() to zero the memory which forces
the pages to be realized and always adds the complete size
to the process's working set.

BUG=612470
Committed: https://crrev.com/eda1a99787e55f9ea71687361e0dd3f54492e603
Cr-Commit-Position: refs/heads/master@{#397489}

[bcwhite, 2016-05-26 20:49:13]

Patchset #2 (id:20001) has been deleted

[bcwhite, 2016-05-27 11:52:14]

Patchset #3 (id:60001) has been deleted

[bcwhite, 2016-05-27 11:54:59]

bcwhite@chromium.org changed reviewers:
+ isherman@chromium.org

[Ilya Sherman, 2016-06-01 23:34:58]

LGTM % comments:

https://codereview.chromium.org/2016013003/diff/80001/base/metrics/persistent...
File base/metrics/persistent_memory_allocator.cc (right):

https://codereview.chromium.org/2016013003/diff/80001/base/metrics/persistent...
base/metrics/persistent_memory_allocator.cc:737: DCHECK(address) <<
::GetLastError();
nit: Mebbe use PCHECK, here and elsewhere where you print GetLastError/errno?

https://codereview.chromium.org/2016013003/diff/80001/base/metrics/persistent...
base/metrics/persistent_memory_allocator.cc:741: // MAP_SHARED is not available
on Linux <2.4 but required on Mac.
Would it make more sense to have separate code paths for Linux and Mac, then?

[bcwhite, 2016-06-02 13:47:05]

https://codereview.chromium.org/2016013003/diff/80001/base/metrics/persistent...
File base/metrics/persistent_memory_allocator.cc (right):

https://codereview.chromium.org/2016013003/diff/80001/base/metrics/persistent...
base/metrics/persistent_memory_allocator.cc:737: DCHECK(address) <<
::GetLastError();
On 2016/06/01 23:34:58, Ilya Sherman wrote:
> nit: Mebbe use PCHECK, here and elsewhere where you print GetLastError/errno?

Done.

https://codereview.chromium.org/2016013003/diff/80001/base/metrics/persistent...
base/metrics/persistent_memory_allocator.cc:741: // MAP_SHARED is not available
on Linux <2.4 but required on Mac.
On 2016/06/01 23:34:58, Ilya Sherman wrote:
> Would it make more sense to have separate code paths for Linux and Mac, then?

I considered that but in the end decided that a combination that worked on both
Mac and Linux was most likely to work on any configuration marked "posix"
without having to individually try every specific OS definition.

[bcwhite, 2016-06-02 15:59:42]

Patchset #4 (id:100001) has been deleted

[bcwhite, 2016-06-02 19:19:55]

The CQ bit was checked by bcwhite@chromium.org

[bcwhite, 2016-06-02 19:19:56]

The patchset sent to the CQ was uploaded after l-g-t-m from
isherman@chromium.org
Link to the patchset: https://codereview.chromium.org/2016013003/#ps120001
(title: "use DPCHECK instead of DCHECK to print system error messages (and
rebased)")

[commit-bot: I haz the power, 2016-06-02 19:21:09]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/2016013003/120001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/2016013003/120001

[commit-bot: I haz the power, 2016-06-02 19:27:32]

Committed patchset #4 (id:120001)

[commit-bot: I haz the power, 2016-06-02 19:29:03]

Description was changed from

==========
Properly create zero-initialized local memory.

Previous code used memset() to zero the memory which forces
the pages to be realized and always adds the complete size
to the process's working set.

BUG=612470
==========

to

==========
Properly create zero-initialized local memory.

Previous code used memset() to zero the memory which forces
the pages to be realized and always adds the complete size
to the process's working set.

BUG=612470

Committed: https://crrev.com/eda1a99787e55f9ea71687361e0dd3f54492e603
Cr-Commit-Position: refs/heads/master@{#397489}
==========

[commit-bot: I haz the power, 2016-06-02 19:29:04]

Patchset 4 (id:??) landed as
https://crrev.com/eda1a99787e55f9ea71687361e0dd3f54492e603
Cr-Commit-Position: refs/heads/master@{#397489}