Consolidate code monitoring low memory kills and OOM kills to MemoryKillsMonitor on ChromeOS.

Consolidate code monitoring low memory kills and OOM kills to MemoryKillsMonitor on ChromeOS

Rename OomKillsMonitor to MemoryKillsMonitor and move it from
arc namespace to memory namespace. Now it is responsible for
1. Log OOM kill events by listening to kernel messages (in a dedicated thread).
2. Log low memory kill events when TabManager kills processes (called from UI thread).

It logs those events to
1. Chrome UMA
2. A local file if --memory-kills-log is given

It starts a new monitoring session when a new BrowserProcess is created.

BUG=none
TEST=manual

Committed: https://crrev.com/43f78af4b909bca7f815b862b9ab3f17c5cacddf
Cr-Commit-Position: refs/heads/master@{#435743}

[cylee1, 2016-11-24 17:47:51]

The CQ bit was checked by cylee@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-24 17:48:13]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[cylee1, 2016-11-24 18:04:12]

cylee@chromium.org changed reviewers:
+ georgesak@chromium.org, hidehiko@chromium.org, oshima@chromium.org,
sky@chromium.org

[cylee1, 2016-11-24 18:04:12]

Add reviewer

sky for files chrome/browser/browser_process_impl.*

oshima for chromeos/chromeos_switches.*

georgesak for files under chrome/browser/memory/

hidehiko for files under components/arc

[commit-bot: I haz the power, 2016-11-24 18:07:02]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-11-24 18:07:03]

Dry run: Try jobs failed on following builders:
  linux_chromium_asan_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[cylee1, 2016-11-24 18:28:10]

The CQ bit was checked by cylee@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-24 18:28:26]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-24 18:46:27]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-11-24 18:46:28]

Dry run: Try jobs failed on following builders:
  linux_chromium_compile_dbg_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)
  mac_chromium_compile_dbg_ng on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_comp...)
  mac_chromium_rel_ng on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_...)

[cylee1, 2016-11-24 19:12:54]

Description was changed from

==========
Consolidate code monitoring low memory kills and OOM kills to
MemoryKillsMonitor.

Rename OomKillsMonitor to MemoryKillsMonitor and move it from
arc namespace to memory namespace. Now it is responsible for
1. Log OOM kill events by listening to kernel messages (in a dedicated thread).
2. Log low memory kill events when TabManager kills processes (called from UI
thread).

It logs those events to
1. Chrome UMA
2. A local file if --memory-kills-log is given

It starts a new monitoring session when a new BrowserProcess is created.

BUG=none
TEST=manual
==========

to

==========
Consolidate code monitoring low memory kills and OOM kills to MemoryKillsMonitor
on ChromeOS

Rename OomKillsMonitor to MemoryKillsMonitor and move it from
arc namespace to memory namespace. Now it is responsible for
1. Log OOM kill events by listening to kernel messages (in a dedicated thread).
2. Log low memory kill events when TabManager kills processes (called from UI
thread).

It logs those events to
1. Chrome UMA
2. A local file if --memory-kills-log is given

It starts a new monitoring session when a new BrowserProcess is created.

BUG=none
TEST=manual
==========

[cylee1, 2016-11-24 19:14:14]

The CQ bit was checked by cylee@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-24 19:14:25]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-24 19:50:33]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-11-24 19:50:34]

Dry run: Try jobs failed on following builders:
  win_chromium_x64_rel_ng on master.tryserver.chromium.win (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_x64_...)

[hidehiko, 2016-11-25 10:13:22]

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/memory/m...
File chrome/browser/memory/memory_kills_histogram.h (right):

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_histogram.h:8: #include
"base/metrics/histogram.h"
#include "base/time/time.h", too?

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_histogram.h:12: const int
kMaxMemoryKillTimeDeltaSecs = 30;
base::TimeDelta supports constexpr. So,

constexpr base::TimeDelta kMaxMemoryKillTimeDelta =
    base::TimeDelta::FromSeconds(30);

seems better?

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/memory/m...
File chrome/browser/memory/memory_kills_monitor.cc (right):

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_monitor.cc:137: void ParseDmesg(const
StringPiece& line) {
This function seems doing more than parsing. (Actually logging).
Could you rename to the appropriate one? Or could you return the parsed value,
and log in the caller function?

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_monitor.cc:173: log_file_ =
base::ScopedFILE(base::OpenFile(log_file_name, "a"));
Is this thread allowed to access files?

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_monitor.cc:197: g_instance = new
MemoryKillsMonitor;
Could you add trailing ()?
cf) 2) of
https://www.chromium.org/developers/coding-style/cpp-dos-and-donts#TOC-Prefer...

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_monitor.cc:238: base::AutoLock
lock(log_file_lock_);
If this is the guard for log_file_, then L235 needs to be guarded, as well?

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/memory/m...
File chrome/browser/memory/memory_kills_monitor.h (right):

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_monitor.h:39: class Handle {
Looks like this is RAII, so should this be non-copyable?

[sky, 2016-11-26 16:49:40]

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/browser_...
File chrome/browser/browser_process_impl.cc (right):

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/browser_...
chrome/browser/browser_process_impl.cc:218: #endif  // !defined(OS_ANDROID)
ANDROID->CHROMEOS

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/browser_...
File chrome/browser/browser_process_impl.h (right):

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/browser_...
chrome/browser/browser_process_impl.h:350: #if defined(OS_CHROMEOS)
If this is CHROMEOS specific, how come the include isn't only for chromeos? That
said, AFAICT this member isn't used by BrowserProcessImpl. Can it move to
ChromeBrowserMainPartsChromeos?

[cylee1, 2016-11-29 20:24:51]

The CQ bit was checked by cylee@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-29 20:25:16]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[cylee1, 2016-11-29 20:28:41]

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/browser_...
File chrome/browser/browser_process_impl.cc (right):

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/browser_...
chrome/browser/browser_process_impl.cc:218: #endif  // !defined(OS_ANDROID)
On 2016/11/26 16:49:40, sky wrote:
> ANDROID->CHROMEOS

Done.

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/browser_...
File chrome/browser/browser_process_impl.h (right):

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/browser_...
chrome/browser/browser_process_impl.h:350: #if defined(OS_CHROMEOS)
On 2016/11/26 16:49:40, sky wrote:
> If this is CHROMEOS specific, how come the include isn't only for chromeos?
That
> said, AFAICT this member isn't used by BrowserProcessImpl. Can it move to
> ChromeBrowserMainPartsChromeos?

You're right. Moved it to ChromeBrowserMainPartsChromeos.
Also only include memory_kills_monitor.[cc.h] in ChromeOS build in BUILD.gn

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/memory/m...
File chrome/browser/memory/memory_kills_histogram.h (right):

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_histogram.h:8: #include
"base/metrics/histogram.h"
On 2016/11/25 10:13:21, hidehiko wrote:
> #include "base/time/time.h", too?

Done.

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_histogram.h:12: const int
kMaxMemoryKillTimeDeltaSecs = 30;
On 2016/11/25 10:13:21, hidehiko wrote:
> base::TimeDelta supports constexpr. So,
> 
> constexpr base::TimeDelta kMaxMemoryKillTimeDelta =
>     base::TimeDelta::FromSeconds(30);
> 
> seems better?

Done.

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/memory/m...
File chrome/browser/memory/memory_kills_monitor.cc (right):

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_monitor.cc:137: void ParseDmesg(const
StringPiece& line) {
On 2016/11/25 10:13:21, hidehiko wrote:
> This function seems doing more than parsing. (Actually logging).
> Could you rename to the appropriate one? Or could you return the parsed value,
> and log in the caller function?

Done.

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_monitor.cc:173: log_file_ =
base::ScopedFILE(base::OpenFile(log_file_name, "a"));
On 2016/11/25 10:13:21, hidehiko wrote:
> Is this thread allowed to access files?

You're right. I decide to log to standard log file with VLOG

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_monitor.cc:197: g_instance = new
MemoryKillsMonitor;
On 2016/11/25 10:13:21, hidehiko wrote:
> Could you add trailing ()?
> cf) 2) of
>
https://www.chromium.org/developers/coding-style/cpp-dos-and-donts#TOC-Prefer...

Done.

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_monitor.cc:238: base::AutoLock
lock(log_file_lock_);
On 2016/11/25 10:13:21, hidehiko wrote:
> If this is the guard for log_file_, then L235 needs to be guarded, as well?

The file lock is removed.

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/memory/m...
File chrome/browser/memory/memory_kills_monitor.h (right):

https://codereview.chromium.org/2527973003/diff/60001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_monitor.h:39: class Handle {
On 2016/11/25 10:13:21, hidehiko wrote:
> Looks like this is RAII, so should this be non-copyable?

Looks like so (It's added by gab@).
It requires a move constructor now (and looks more reasonable).

[sky, 2016-11-29 20:55:02]

hrome/browser/chromeos/chrome_browser_main_chromeos.* LGTM

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/chromeos...
File chrome/browser/chromeos/chrome_browser_main_chromeos.cc (right):

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/chromeos...
chrome/browser/chromeos/chrome_browser_main_chromeos.cc:358:
memory_kills_monitor_.reset(new memory::MemoryKillsMonitor::Handle(
Use MakeUnique if you can (see threads on chromium-dev).

[commit-bot: I haz the power, 2016-11-29 22:26:16]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-11-29 22:26:18]

Dry run: This issue passed the CQ dry run.

[hidehiko, 2016-11-30 09:05:49]

Looks good except Handle move ctor.
Others are minor comments.

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/memory/m...
File chrome/browser/memory/memory_kills_monitor.cc (right):

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_monitor.cc:55: void LogEvent(base::Time
time_stamp, const std::string& event) {
Could you use "const base::Time&" instead?

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_monitor.cc:196: static const int kMaxBufSize
= 512;
s/static const/constexpr/

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_monitor.cc:199: while (fgets(buf,
kMaxBufSize, kmsg_handle.get())) {
Slightly off topic:
I think this is out of focus of this CL, so I'm ok to leave it as is... but
please let me note.

fgets() is a blocking call, so on shutdown is_shutting_down_.IsSet() check will
be pending until /dev/kmsg receives new message. Is it ensured? If so, maybe
better to comment? If not, we cannot shutdown gracefully?

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_monitor.cc:204: const StringPiece
buf_string(buf);
Optional:

How about,

void TryMatchOomKillLine(const std::string& line)

so that re2::StringPiece can directly take the line, instead of data() and
length().

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/memory/m...
File chrome/browser/memory/memory_kills_monitor.h (right):

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_monitor.h:46: Handle(Handle&& handle) =
default;
Does this actually work? IIUC, ptr move is just a copy?

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/memory/t...
File chrome/browser/memory/tab_manager_delegate_chromeos.cc (right):

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/memory/t...
chrome/browser/memory/tab_manager_delegate_chromeos.cc:27: #include
"base/time/time.h"
Unused now?

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/memory/t...
chrome/browser/memory/tab_manager_delegate_chromeos.cc:51: using
base::TimeDelta;
Unused?

[cylee1, 2016-11-30 17:18:58]

Ping Georges

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/chromeos...
File chrome/browser/chromeos/chrome_browser_main_chromeos.cc (right):

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/chromeos...
chrome/browser/chromeos/chrome_browser_main_chromeos.cc:358:
memory_kills_monitor_.reset(new memory::MemoryKillsMonitor::Handle(
On 2016/11/29 20:55:02, sky wrote:
> Use MakeUnique if you can (see threads on chromium-dev).

Done.

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/memory/m...
File chrome/browser/memory/memory_kills_monitor.cc (right):

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_monitor.cc:55: void LogEvent(base::Time
time_stamp, const std::string& event) {
On 2016/11/30 09:05:48, hidehiko wrote:
> Could you use "const base::Time&" instead?

Done.

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_monitor.cc:196: static const int kMaxBufSize
= 512;
On 2016/11/30 09:05:48, hidehiko wrote:
> s/static const/constexpr/

Done.

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_monitor.cc:199: while (fgets(buf,
kMaxBufSize, kmsg_handle.get())) {
On 2016/11/30 09:05:48, hidehiko wrote:
> Slightly off topic:
> I think this is out of focus of this CL, so I'm ok to leave it as is... but
> please let me note.
> 
> fgets() is a blocking call, so on shutdown is_shutting_down_.IsSet() check
will
> be pending until /dev/kmsg receives new message. Is it ensured? If so, maybe
> better to comment? If not, we cannot shutdown gracefully?

Yeah that's what bugs me. Because I want to handle a line at a time for
simplicity, I have to use line buffered blocking calls.
It's also the reason why it needs to be in a non-joinable thread. It didn't
block the process from shutting down though. 
I tried to find a way to solve the problem you raised when I design this, but
didn't find a good solution.

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_monitor.cc:204: const StringPiece
buf_string(buf);
On 2016/11/30 09:05:48, hidehiko wrote:
> Optional:
> 
> How about,
> 
> void TryMatchOomKillLine(const std::string& line)
> 
> so that re2::StringPiece can directly take the line, instead of data() and
> length().

Done.

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/memory/m...
File chrome/browser/memory/memory_kills_monitor.h (right):

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/memory/m...
chrome/browser/memory/memory_kills_monitor.h:46: Handle(Handle&& handle) =
default;
On 2016/11/30 09:05:48, hidehiko wrote:
> Does this actually work? IIUC, ptr move is just a copy?

Re-implemented. I should nullify the moved instance.

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/memory/t...
File chrome/browser/memory/tab_manager_delegate_chromeos.cc (right):

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/memory/t...
chrome/browser/memory/tab_manager_delegate_chromeos.cc:27: #include
"base/time/time.h"
On 2016/11/30 09:05:48, hidehiko wrote:
> Unused now?

still used

https://codereview.chromium.org/2527973003/diff/80001/chrome/browser/memory/t...
chrome/browser/memory/tab_manager_delegate_chromeos.cc:51: using
base::TimeDelta;
On 2016/11/30 09:05:48, hidehiko wrote:
> Unused?

still used (e.g., line 343)

[Georges Khalil, 2016-11-30 20:53:35]

chrome/browser/memory LGTM % nit

But shouldn't we have tests for the functions in memory_kills_monitor.cc?

https://codereview.chromium.org/2527973003/diff/120001/chrome/browser/memory/...
File chrome/browser/memory/memory_kills_monitor.h (right):

https://codereview.chromium.org/2527973003/diff/120001/chrome/browser/memory/...
chrome/browser/memory/memory_kills_monitor.h:22: // TabManagr).
nit: TabManager.

[hidehiko, 2016-12-01 12:40:08]

components/arc LGTM.

[cylee1, 2016-12-01 20:02:45]

Hi Georges,
  Unit test added.

https://codereview.chromium.org/2527973003/diff/120001/chrome/browser/memory/...
File chrome/browser/memory/memory_kills_monitor.h (right):

https://codereview.chromium.org/2527973003/diff/120001/chrome/browser/memory/...
chrome/browser/memory/memory_kills_monitor.h:22: // TabManagr).
On 2016/11/30 20:53:35, Georges Khalil wrote:
> nit: TabManager.

Done.

[Georges Khalil, 2016-12-01 20:45:46]

Still LGTM.

Thanks for the tests!

[cylee1, 2016-12-01 20:53:23]

The CQ bit was checked by cylee@chromium.org

[cylee1, 2016-12-01 20:53:24]

The patchset sent to the CQ was uploaded after l-g-t-m from sky@chromium.org,
hidehiko@chromium.org
Link to the patchset: https://codereview.chromium.org/2527973003/#ps140001
(title: "Add unittest")

[commit-bot: I haz the power, 2016-12-01 20:54:18]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-12-01 21:14:58]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-12-01 21:14:59]

Try jobs failed on following builders:
  mac_chromium_rel_ng on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_...)

[cylee1, 2016-12-01 21:38:19]

The CQ bit was checked by cylee@chromium.org

[cylee1, 2016-12-01 21:38:19]

The patchset sent to the CQ was uploaded after l-g-t-m from
hidehiko@chromium.org, georgesak@chromium.org, sky@chromium.org
Link to the patchset: https://codereview.chromium.org/2527973003/#ps160001
(title: "move unittest to chromeos only build rule")

[commit-bot: I haz the power, 2016-12-01 21:38:53]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-12-01 22:43:55]

CQ is committing da patch.

Bot data: {"patchset_id": 160001, "attempt_start_ts": 1480628299416740,
"parent_rev": "4c0b0447635fb7435a54925cc48545af14455c14", "commit_rev":
"e1404fbd0bede593bd3c082dafe7d26445ff995a"}

[commit-bot: I haz the power, 2016-12-01 22:44:22]

Description was changed from

==========
Consolidate code monitoring low memory kills and OOM kills to MemoryKillsMonitor
on ChromeOS

Rename OomKillsMonitor to MemoryKillsMonitor and move it from
arc namespace to memory namespace. Now it is responsible for
1. Log OOM kill events by listening to kernel messages (in a dedicated thread).
2. Log low memory kill events when TabManager kills processes (called from UI
thread).

It logs those events to
1. Chrome UMA
2. A local file if --memory-kills-log is given

It starts a new monitoring session when a new BrowserProcess is created.

BUG=none
TEST=manual
==========

to

==========
Consolidate code monitoring low memory kills and OOM kills to MemoryKillsMonitor
on ChromeOS

Rename OomKillsMonitor to MemoryKillsMonitor and move it from
arc namespace to memory namespace. Now it is responsible for
1. Log OOM kill events by listening to kernel messages (in a dedicated thread).
2. Log low memory kill events when TabManager kills processes (called from UI
thread).

It logs those events to
1. Chrome UMA
2. A local file if --memory-kills-log is given

It starts a new monitoring session when a new BrowserProcess is created.

BUG=none
TEST=manual
==========

[commit-bot: I haz the power, 2016-12-01 22:44:23]

Committed patchset #9 (id:160001)

[commit-bot: I haz the power, 2016-12-01 22:46:10]

Description was changed from

==========
Consolidate code monitoring low memory kills and OOM kills to MemoryKillsMonitor
on ChromeOS

Rename OomKillsMonitor to MemoryKillsMonitor and move it from
arc namespace to memory namespace. Now it is responsible for
1. Log OOM kill events by listening to kernel messages (in a dedicated thread).
2. Log low memory kill events when TabManager kills processes (called from UI
thread).

It logs those events to
1. Chrome UMA
2. A local file if --memory-kills-log is given

It starts a new monitoring session when a new BrowserProcess is created.

BUG=none
TEST=manual
==========

to

==========
Consolidate code monitoring low memory kills and OOM kills to MemoryKillsMonitor
on ChromeOS

Rename OomKillsMonitor to MemoryKillsMonitor and move it from
arc namespace to memory namespace. Now it is responsible for
1. Log OOM kill events by listening to kernel messages (in a dedicated thread).
2. Log low memory kill events when TabManager kills processes (called from UI
thread).

It logs those events to
1. Chrome UMA
2. A local file if --memory-kills-log is given

It starts a new monitoring session when a new BrowserProcess is created.

BUG=none
TEST=manual

Committed: https://crrev.com/43f78af4b909bca7f815b862b9ab3f17c5cacddf
Cr-Commit-Position: refs/heads/master@{#435743}
==========

[commit-bot: I haz the power, 2016-12-01 22:46:11]

Patchset 9 (id:??) landed as
https://crrev.com/43f78af4b909bca7f815b862b9ab3f17c5cacddf
Cr-Commit-Position: refs/heads/master@{#435743}