Experimental 'purging and suspending' backgrounded tabs behind the flag

Experimental 'purging and suspending' backgrounded tabs behind the flag

This CL adds an experimental feature, 'purging and suspending'
backgrounded tabs to save memory usage. This feature is behind the
flag '--purge-and-suspend-time' taking a time value in seconds. When
the indicated time passes after a tab goes backgrounded, the
backgrounded tab's memory cache is purged and suspended. We expect
only purgable memory like cache is purged and tab behaviors would never
be affected.

This CL adds only suspending and not purging any cache. I'll add
purging feature later.

BUG=551995, 607077
Committed: https://crrev.com/6909327ee89d7bd5b69e9e578bcb28388271b2ea
Cr-Commit-Position: refs/heads/master@{#393476}

[hajimehoshi, 2016-04-27 07:15:21]

Description was changed from

==========
Experimental 'purging and suspending' backgrounded tabs behind the flag

(Work in progress)

This CL adds an experimental feature, 'purging and suspending'
backgrounded tabs to save memory usage. This features is behind the
flag '--purge-and-suspend-time' taking a time value in seconds. When
the indicated time passes after a tabs goes backgrounded, the
backgrounded tab's memory cache is purged and suspended. We expect
only purgable memory like cache is purged and tab behaviors would never
be affected.

This CL adds only suspending and not purging any cahce. I'll add
purging feature later.

BUG=
==========

to

==========
Experimental 'purging and suspending' backgrounded tabs behind the flag

(Work in progress)

This CL adds an experimental feature, 'purging and suspending'
backgrounded tabs to save memory usage. This features is behind the
flag '--purge-and-suspend-time' taking a time value in seconds. When
the indicated time passes after a tabs goes backgrounded, the
backgrounded tab's memory cache is purged and suspended. We expect
only purgable memory like cache is purged and tab behaviors would never
be affected.

This CL adds only suspending and not purging any cahce. I'll add
purging feature later.

BUG=551995, 607077
==========

[hajimehoshi, 2016-04-27 08:37:10]

Description was changed from

==========
Experimental 'purging and suspending' backgrounded tabs behind the flag

(Work in progress)

This CL adds an experimental feature, 'purging and suspending'
backgrounded tabs to save memory usage. This features is behind the
flag '--purge-and-suspend-time' taking a time value in seconds. When
the indicated time passes after a tabs goes backgrounded, the
backgrounded tab's memory cache is purged and suspended. We expect
only purgable memory like cache is purged and tab behaviors would never
be affected.

This CL adds only suspending and not purging any cahce. I'll add
purging feature later.

BUG=551995, 607077
==========

to

==========
Experimental 'purging and suspending' backgrounded tabs behind the flag

This CL adds an experimental feature, 'purging and suspending'
backgrounded tabs to save memory usage. This features is behind the
flag '--purge-and-suspend-time' taking a time value in seconds. When
the indicated time passes after a tabs goes backgrounded, the
backgrounded tab's memory cache is purged and suspended. We expect
only purgable memory like cache is purged and tab behaviors would never
be affected.

This CL adds only suspending and not purging any cahce. I'll add
purging feature later.

BUG=551995, 607077
==========

[hajimehoshi, 2016-04-27 08:38:17]

hajimehoshi@chromium.org changed reviewers:
+ chrisha@chromium.org, georgesak@chromium.org, haraken@chromium.org

[hajimehoshi, 2016-04-27 08:38:18]

PTAL

[Sami, 2016-04-27 15:44:42]

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

[Sami, 2016-04-27 15:44:43]

https://codereview.chromium.org/1914143002/diff/40001/components/scheduler/re...
File components/scheduler/renderer/renderer_scheduler_impl.cc (right):

https://codereview.chromium.org/1914143002/diff/40001/components/scheduler/re...
components/scheduler/renderer/renderer_scheduler_impl.cc:389: if
(helper_.IsShutdown() || !MainThreadOnly().renderer_backgrounded)
We should probably cancel suspend_timers_when_backgrounded_closure_ here because
this is doing essentially the same thing.

https://codereview.chromium.org/1914143002/diff/40001/components/scheduler/re...
components/scheduler/renderer/renderer_scheduler_impl.cc:391: //
TODO(hajimehoshi): Implement purging e.g. cache (crbug/607077)
Have you thought about what kind of interfaces would you need to implement
purging? Since this is in components it might be a little awkward. (One proposal
has been to move RendererScheduler to blink/platform, but so far we haven't
investigated it too deeply.)

https://codereview.chromium.org/1914143002/diff/40001/components/scheduler/re...
components/scheduler/renderer/renderer_scheduler_impl.cc:392:
SuspendTimerQueueWhenBackgrounded();
We should think about whether we need to suspend other task sources such as
network tasks and perhaps postMessages.

[hajimehoshi, 2016-04-28 10:26:58]

Thank you!

https://codereview.chromium.org/1914143002/diff/40001/components/scheduler/re...
File components/scheduler/renderer/renderer_scheduler_impl.cc (right):

https://codereview.chromium.org/1914143002/diff/40001/components/scheduler/re...
components/scheduler/renderer/renderer_scheduler_impl.cc:389: if
(helper_.IsShutdown() || !MainThreadOnly().renderer_backgrounded)
On 2016/04/27 15:44:43, Sami (slow -- travelling) wrote:
> We should probably cancel suspend_timers_when_backgrounded_closure_ here
because
> this is doing essentially the same thing.

Done.

https://codereview.chromium.org/1914143002/diff/40001/components/scheduler/re...
components/scheduler/renderer/renderer_scheduler_impl.cc:391: //
TODO(hajimehoshi): Implement purging e.g. cache (crbug/607077)
On 2016/04/27 15:44:43, Sami (slow -- travelling) wrote:
> Have you thought about what kind of interfaces would you need to implement
> purging? Since this is in components it might be a little awkward. (One
proposal
> has been to move RendererScheduler to blink/platform, but so far we haven't
> investigated it too deeply.)

Right. Moved this comment to RenderThreadImpl and changed the scheduler to do
only suspending. Done.

https://codereview.chromium.org/1914143002/diff/40001/components/scheduler/re...
components/scheduler/renderer/renderer_scheduler_impl.cc:392:
SuspendTimerQueueWhenBackgrounded();
Good point, we might need to do what you suggest. However, we expect suspending
only the timer queue will stop not all but most of layout and JavaScript. I'll
revisit this when needed since this is just an experiment.

[Sami, 2016-04-28 13:16:28]

Thanks, I think this division of work seems appropriate. Please add a
RendererSchedulerImpl test for this once this graduates into an actual
intervention :)

[hajimehoshi, 2016-05-09 08:58:34]

On 2016/04/28 13:16:28, Sami wrote:
> Thanks, I think this division of work seems appropriate. Please add a
> RendererSchedulerImpl test for this once this graduates into an actual
> intervention :)

Thanks, added a test. Done.

[hajimehoshi, 2016-05-09 12:13:22]

hajimehoshi@chromium.org changed reviewers:
+ kinuko@chromium.org

[hajimehoshi, 2016-05-09 12:16:31]

hajimehoshi@chromium.org changed reviewers:
+ dcheng@chromium.org

[hajimehoshi, 2016-05-09 12:16:56]

+kinuko for content/
+dcheng for content/common

[dcheng, 2016-05-10 04:49:54]

https://codereview.chromium.org/1914143002/diff/120001/chrome/browser/memory/...
File chrome/browser/memory/tab_manager.cc (right):

https://codereview.chromium.org/1914143002/diff/120001/chrome/browser/memory/...
chrome/browser/memory/tab_manager.cc:571: if (NowTicks() - tab.last_active <
purge_and_suspend_time_delta)
Nit: maybe just get NowTicks() once outside the loop and then just do a simple
comparison in here? e.g.

auto purge_and_suspend_threshhold  = NowTicks() -
base::TimeDelta::FromSeconds(purge_and_suspend_time);

for (...) {
  if (tab.last_active < purge_and_suspend_threshhold)
  /* ... */
}

[hajimehoshi, 2016-05-10 06:54:11]

Thank you!

https://codereview.chromium.org/1914143002/diff/120001/chrome/browser/memory/...
File chrome/browser/memory/tab_manager.cc (right):

https://codereview.chromium.org/1914143002/diff/120001/chrome/browser/memory/...
chrome/browser/memory/tab_manager.cc:571: if (NowTicks() - tab.last_active <
purge_and_suspend_time_delta)
On 2016/05/10 04:49:53, dcheng wrote:
> Nit: maybe just get NowTicks() once outside the loop and then just do a simple
> comparison in here? e.g.
> 
> auto purge_and_suspend_threshhold  = NowTicks() -
> base::TimeDelta::FromSeconds(purge_and_suspend_time);
> 
> for (...) {
>   if (tab.last_active < purge_and_suspend_threshhold)
>   /* ... */
> }

Done.

[kinuko, 2016-05-10 10:42:09]

content lgtm

[dcheng, 2016-05-10 16:28:06]

https://codereview.chromium.org/1914143002/diff/140001/content/renderer/rende...
File content/renderer/render_thread_impl.cc (right):

https://codereview.chromium.org/1914143002/diff/140001/content/renderer/rende...
content/renderer/render_thread_impl.cc:1698:
renderer_scheduler_->SuspendRendererWhenBackgrounded();
Will we ever send a "purge and suspend" to a non-background renderer?

[Georges Khalil, 2016-05-10 18:58:58]

In Description:
- This features is -> This feature is
- after a tabs goes -> after a tab goes
- not purging any cahce -> not purging any cache

https://codereview.chromium.org/1914143002/diff/140001/chrome/browser/memory/...
File chrome/browser/memory/tab_manager.cc (right):

https://codereview.chromium.org/1914143002/diff/140001/chrome/browser/memory/...
chrome/browser/memory/tab_manager.cc:518: // updating on ChromeOS via the
delegate.
nit: Add a mention about purging and suspending.

https://codereview.chromium.org/1914143002/diff/140001/chrome/browser/memory/...
File chrome/browser/memory/tab_manager_browsertest.cc (right):

https://codereview.chromium.org/1914143002/diff/140001/chrome/browser/memory/...
chrome/browser/memory/tab_manager_browsertest.cc:46:
command_line->AppendSwitchASCII(switches::kPurgeAndSuspendTime, "1");
Do we need this here?

[hajimehoshi, 2016-05-11 06:23:12]

Description was changed from

==========
Experimental 'purging and suspending' backgrounded tabs behind the flag

This CL adds an experimental feature, 'purging and suspending'
backgrounded tabs to save memory usage. This features is behind the
flag '--purge-and-suspend-time' taking a time value in seconds. When
the indicated time passes after a tabs goes backgrounded, the
backgrounded tab's memory cache is purged and suspended. We expect
only purgable memory like cache is purged and tab behaviors would never
be affected.

This CL adds only suspending and not purging any cahce. I'll add
purging feature later.

BUG=551995, 607077
==========

to

==========
Experimental 'purging and suspending' backgrounded tabs behind the flag

This CL adds an experimental feature, 'purging and suspending'
backgrounded tabs to save memory usage. This feature is behind the
flag '--purge-and-suspend-time' taking a time value in seconds. When
the indicated time passes after a tabs goes backgrounded, the
backgrounded tab's memory cache is purged and suspended. We expect
only purgable memory like cache is purged and tab behaviors would never
be affected.

This CL adds only suspending and not purging any cahce. I'll add
purging feature later.

BUG=551995, 607077
==========

[hajimehoshi, 2016-05-11 07:22:25]

Description was changed from

==========
Experimental 'purging and suspending' backgrounded tabs behind the flag

This CL adds an experimental feature, 'purging and suspending'
backgrounded tabs to save memory usage. This feature is behind the
flag '--purge-and-suspend-time' taking a time value in seconds. When
the indicated time passes after a tabs goes backgrounded, the
backgrounded tab's memory cache is purged and suspended. We expect
only purgable memory like cache is purged and tab behaviors would never
be affected.

This CL adds only suspending and not purging any cahce. I'll add
purging feature later.

BUG=551995, 607077
==========

to

==========
Experimental 'purging and suspending' backgrounded tabs behind the flag

This CL adds an experimental feature, 'purging and suspending'
backgrounded tabs to save memory usage. This feature is behind the
flag '--purge-and-suspend-time' taking a time value in seconds. When
the indicated time passes after a tab goes backgrounded, the
backgrounded tab's memory cache is purged and suspended. We expect
only purgable memory like cache is purged and tab behaviors would never
be affected.

This CL adds only suspending and not purging any cahce. I'll add
purging feature later.

BUG=551995, 607077
==========

[hajimehoshi, 2016-05-11 07:23:02]

Description was changed from

==========
Experimental 'purging and suspending' backgrounded tabs behind the flag

This CL adds an experimental feature, 'purging and suspending'
backgrounded tabs to save memory usage. This feature is behind the
flag '--purge-and-suspend-time' taking a time value in seconds. When
the indicated time passes after a tab goes backgrounded, the
backgrounded tab's memory cache is purged and suspended. We expect
only purgable memory like cache is purged and tab behaviors would never
be affected.

This CL adds only suspending and not purging any cahce. I'll add
purging feature later.

BUG=551995, 607077
==========

to

==========
Experimental 'purging and suspending' backgrounded tabs behind the flag

This CL adds an experimental feature, 'purging and suspending'
backgrounded tabs to save memory usage. This feature is behind the
flag '--purge-and-suspend-time' taking a time value in seconds. When
the indicated time passes after a tab goes backgrounded, the
backgrounded tab's memory cache is purged and suspended. We expect
only purgable memory like cache is purged and tab behaviors would never
be affected.

This CL adds only suspending and not purging any cache. I'll add
purging feature later.

BUG=551995, 607077
==========

[hajimehoshi, 2016-05-11 07:23:31]

Thank you!

https://codereview.chromium.org/1914143002/diff/140001/chrome/browser/memory/...
File chrome/browser/memory/tab_manager.cc (right):

https://codereview.chromium.org/1914143002/diff/140001/chrome/browser/memory/...
chrome/browser/memory/tab_manager.cc:518: // updating on ChromeOS via the
delegate.
On 2016/05/10 18:58:58, Georges Khalil wrote:
> nit: Add a mention about purging and suspending.

Done.

https://codereview.chromium.org/1914143002/diff/140001/chrome/browser/memory/...
File chrome/browser/memory/tab_manager_browsertest.cc (right):

https://codereview.chromium.org/1914143002/diff/140001/chrome/browser/memory/...
chrome/browser/memory/tab_manager_browsertest.cc:46:
command_line->AppendSwitchASCII(switches::kPurgeAndSuspendTime, "1");
On 2016/05/10 18:58:58, Georges Khalil wrote:
> Do we need this here?

No, Done.

https://codereview.chromium.org/1914143002/diff/140001/content/renderer/rende...
File content/renderer/render_thread_impl.cc (right):

https://codereview.chromium.org/1914143002/diff/140001/content/renderer/rende...
content/renderer/render_thread_impl.cc:1698:
renderer_scheduler_->SuspendRendererWhenBackgrounded();
On 2016/05/10 16:28:06, dcheng wrote:
> Will we ever send a "purge and suspend" to a non-background renderer?

Good point. Done by adding |is_backgrounded_|.

[hajimehoshi, 2016-05-11 07:23:49]

On 2016/05/10 18:58:58, Georges Khalil wrote:
> In Description:
> - This features is -> This feature is
> - after a tabs goes -> after a tab goes
> - not purging any cahce -> not purging any cache

Done.

[Georges Khalil, 2016-05-11 15:01:23]

chrome/browser/memory/* lgtm % nit.

https://codereview.chromium.org/1914143002/diff/160001/chrome/browser/memory/...
File chrome/browser/memory/tab_manager_browsertest.cc (right):

https://codereview.chromium.org/1914143002/diff/160001/chrome/browser/memory/...
chrome/browser/memory/tab_manager_browsertest.cc:20: #include
"chrome/common/chrome_switches.h"
nit: remove this.

[dcheng, 2016-05-11 17:42:22]

https://codereview.chromium.org/1914143002/diff/160001/content/renderer/rende...
File content/renderer/render_thread_impl.cc (right):

https://codereview.chromium.org/1914143002/diff/160001/content/renderer/rende...
content/renderer/render_thread_impl.cc:1702:
renderer_scheduler_->SuspendRendererWhenBackgrounded();
Sorry, to clarify my earlier comment: it seems like in task_manager.cc, we'll
never send this IPC if the renderer is still foreground. So why do we need to
call this? If we get the purge and suspend IPC message, won't the renderer
already have been backgrounded by the ChildProcessMsg_SetProcessBackgrounded
message?

[hajimehoshi, 2016-05-12 05:20:14]

https://codereview.chromium.org/1914143002/diff/160001/content/renderer/rende...
File content/renderer/render_thread_impl.cc (right):

https://codereview.chromium.org/1914143002/diff/160001/content/renderer/rende...
content/renderer/render_thread_impl.cc:1702:
renderer_scheduler_->SuspendRendererWhenBackgrounded();
On 2016/05/11 17:42:22, dcheng wrote:
> Sorry, to clarify my earlier comment: it seems like in task_manager.cc, we'll
> never send this IPC if the renderer is still foreground. So why do we need to
> call this? If we get the purge and suspend IPC message, won't the renderer
> already have been backgrounded by the ChildProcessMsg_SetProcessBackgrounded
> message?

I understood that checking |is_backgrounded| is not needed here, but assuming
the renderer is backgrounded, SuspendRendererWhenBackgrounded is still needed to
stop the timer queue in the renderer to prevent the return of purged cache
memory.

I was not sure if a tab becomes foregrounded again (e.g. when another tab
sharing the renderer is activated) before the purging IPC arrives to the
renderer, but this never happens, right?

[dcheng, 2016-05-12 05:25:47]

On 2016/05/12 at 05:20:14, hajimehoshi wrote:
>
https://codereview.chromium.org/1914143002/diff/160001/content/renderer/rende...
> File content/renderer/render_thread_impl.cc (right):
> 
>
https://codereview.chromium.org/1914143002/diff/160001/content/renderer/rende...
> content/renderer/render_thread_impl.cc:1702:
renderer_scheduler_->SuspendRendererWhenBackgrounded();
> On 2016/05/11 17:42:22, dcheng wrote:
> > Sorry, to clarify my earlier comment: it seems like in task_manager.cc,
we'll
> > never send this IPC if the renderer is still foreground. So why do we need
to
> > call this? If we get the purge and suspend IPC message, won't the renderer
> > already have been backgrounded by the ChildProcessMsg_SetProcessBackgrounded
> > message?
> 
> I understood that checking |is_backgrounded| is not needed here, but assuming
the renderer is backgrounded, SuspendRendererWhenBackgrounded is still needed to
stop the timer queue in the renderer to prevent the return of purged cache
memory.
> 
> I was not sure if a tab becomes foregrounded again (e.g. when another tab
sharing the renderer is activated) before the purging IPC arrives to the
renderer, but this never happens, right?

Assuming that all the browser-side work (to make renderers foreground/background
and to determine if something should be purged+suspended) is happening on the
same thread (UI thread?), the IPC messages will be processed in FIFO order. So
it would be a logic bug for a renderer to be foregrounded again before the
purging IPC is processed.

[hajimehoshi, 2016-05-12 08:33:18]

Thank you!

https://codereview.chromium.org/1914143002/diff/160001/chrome/browser/memory/...
File chrome/browser/memory/tab_manager_browsertest.cc (right):

https://codereview.chromium.org/1914143002/diff/160001/chrome/browser/memory/...
chrome/browser/memory/tab_manager_browsertest.cc:20: #include
"chrome/common/chrome_switches.h"
On 2016/05/11 15:01:23, Georges Khalil wrote:
> nit: remove this.

Done.

https://codereview.chromium.org/1914143002/diff/160001/content/renderer/rende...
File content/renderer/render_thread_impl.cc (right):

https://codereview.chromium.org/1914143002/diff/160001/content/renderer/rende...
content/renderer/render_thread_impl.cc:1702:
renderer_scheduler_->SuspendRendererWhenBackgrounded();
OK, renamed SuspendRendererWhenBackgrounded to SuspendRenderer to clarify the
purpose. also removed |is_backgrounded_|.

[Sami, 2016-05-12 11:31:35]

components/renderer lgtm with two nits.

https://codereview.chromium.org/1914143002/diff/180001/components/scheduler/r...
File components/scheduler/renderer/renderer_scheduler.h (right):

https://codereview.chromium.org/1914143002/diff/180001/components/scheduler/r...
components/scheduler/renderer/renderer_scheduler.h:131: // Tells the scheduler
that the render process should be suspended.
Please also add:

This can only be done when the renderer is backgrounded. The renderer will be
automatically resumed when foregrounded.

https://codereview.chromium.org/1914143002/diff/180001/components/scheduler/r...
File components/scheduler/renderer/renderer_scheduler_impl.cc (right):

https://codereview.chromium.org/1914143002/diff/180001/components/scheduler/r...
components/scheduler/renderer/renderer_scheduler_impl.cc:395:
suspend_timers_when_backgrounded_closure_.Cancel();
Could you add a TODO about possibly suspending other types of tasks too (e.g.,
loading tasks)?

[hajimehoshi, 2016-05-13 05:57:11]

>
https://codereview.chromium.org/1914143002/diff/40001/components/scheduler/re...
> components/scheduler/renderer/renderer_scheduler_impl.cc:392:
> SuspendTimerQueueWhenBackgrounded();
> We should think about whether we need to suspend other task sources such as
> network tasks and perhaps postMessages.

Sami, I realized I didn't know what exactly network tasks mean here. Which class
did you indicate here?

[hajimehoshi, 2016-05-13 06:07:46]

Thank you!

https://codereview.chromium.org/1914143002/diff/180001/components/scheduler/r...
File components/scheduler/renderer/renderer_scheduler.h (right):

https://codereview.chromium.org/1914143002/diff/180001/components/scheduler/r...
components/scheduler/renderer/renderer_scheduler.h:131: // Tells the scheduler
that the render process should be suspended.
On 2016/05/12 11:31:35, Sami wrote:
> Please also add:
> 
> This can only be done when the renderer is backgrounded. The renderer will be
> automatically resumed when foregrounded.

Done.

https://codereview.chromium.org/1914143002/diff/180001/components/scheduler/r...
File components/scheduler/renderer/renderer_scheduler_impl.cc (right):

https://codereview.chromium.org/1914143002/diff/180001/components/scheduler/r...
components/scheduler/renderer/renderer_scheduler_impl.cc:395:
suspend_timers_when_backgrounded_closure_.Cancel();
On 2016/05/12 11:31:35, Sami wrote:
> Could you add a TODO about possibly suspending other types of tasks too (e.g.,
> loading tasks)?

Done.

[dcheng, 2016-05-13 06:08:51]

ipc lgtm

[hajimehoshi, 2016-05-13 06:36:07]

The CQ bit was checked by hajimehoshi@chromium.org

[hajimehoshi, 2016-05-13 06:36:09]

The patchset sent to the CQ was uploaded after l-g-t-m from kinuko@chromium.org,
georgesak@chromium.org, skyostil@chromium.org, dcheng@chromium.org
Link to the patchset: https://codereview.chromium.org/1914143002/#ps220001
(title: "(rebasing)")

[commit-bot: I haz the power, 2016-05-13 06:38:25]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1914143002/220001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1914143002/220001

[haraken, 2016-05-13 07:23:32]

LGTM

https://codereview.chromium.org/1914143002/diff/220001/content/renderer/rende...
File content/renderer/render_thread_impl.cc (right):

https://codereview.chromium.org/1914143002/diff/220001/content/renderer/rende...
content/renderer/render_thread_impl.cc:1725: is_cache_purged_ = true;

Nit: It would make more sense to rename this to is_renderer_suspended.

[hajimehoshi, 2016-05-13 07:31:30]

The CQ bit was unchecked by hajimehoshi@chromium.org

[hajimehoshi, 2016-05-13 07:36:05]

Thank you!

https://codereview.chromium.org/1914143002/diff/220001/content/renderer/rende...
File content/renderer/render_thread_impl.cc (right):

https://codereview.chromium.org/1914143002/diff/220001/content/renderer/rende...
content/renderer/render_thread_impl.cc:1725: is_cache_purged_ = true;
On 2016/05/13 07:23:32, haraken wrote:
> 
> Nit: It would make more sense to rename this to is_renderer_suspended.

Done.

[hajimehoshi, 2016-05-13 07:36:11]

The CQ bit was checked by hajimehoshi@chromium.org

[hajimehoshi, 2016-05-13 07:36:12]

The patchset sent to the CQ was uploaded after l-g-t-m from
skyostil@chromium.org, georgesak@chromium.org, dcheng@chromium.org,
kinuko@chromium.org, haraken@chromium.org
Link to the patchset: https://codereview.chromium.org/1914143002/#ps240001
(title: "Address on haraken's review")

[commit-bot: I haz the power, 2016-05-13 07:36:28]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1914143002/240001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1914143002/240001

[commit-bot: I haz the power, 2016-05-13 08:31:08]

Description was changed from

==========
Experimental 'purging and suspending' backgrounded tabs behind the flag

This CL adds an experimental feature, 'purging and suspending'
backgrounded tabs to save memory usage. This feature is behind the
flag '--purge-and-suspend-time' taking a time value in seconds. When
the indicated time passes after a tab goes backgrounded, the
backgrounded tab's memory cache is purged and suspended. We expect
only purgable memory like cache is purged and tab behaviors would never
be affected.

This CL adds only suspending and not purging any cache. I'll add
purging feature later.

BUG=551995, 607077
==========

to

==========
Experimental 'purging and suspending' backgrounded tabs behind the flag

This CL adds an experimental feature, 'purging and suspending'
backgrounded tabs to save memory usage. This feature is behind the
flag '--purge-and-suspend-time' taking a time value in seconds. When
the indicated time passes after a tab goes backgrounded, the
backgrounded tab's memory cache is purged and suspended. We expect
only purgable memory like cache is purged and tab behaviors would never
be affected.

This CL adds only suspending and not purging any cache. I'll add
purging feature later.

BUG=551995, 607077
==========

[commit-bot: I haz the power, 2016-05-13 08:31:09]

Committed patchset #13 (id:240001)

[commit-bot: I haz the power, 2016-05-13 08:32:49]

Description was changed from

==========
Experimental 'purging and suspending' backgrounded tabs behind the flag

This CL adds an experimental feature, 'purging and suspending'
backgrounded tabs to save memory usage. This feature is behind the
flag '--purge-and-suspend-time' taking a time value in seconds. When
the indicated time passes after a tab goes backgrounded, the
backgrounded tab's memory cache is purged and suspended. We expect
only purgable memory like cache is purged and tab behaviors would never
be affected.

This CL adds only suspending and not purging any cache. I'll add
purging feature later.

BUG=551995, 607077
==========

to

==========
Experimental 'purging and suspending' backgrounded tabs behind the flag

This CL adds an experimental feature, 'purging and suspending'
backgrounded tabs to save memory usage. This feature is behind the
flag '--purge-and-suspend-time' taking a time value in seconds. When
the indicated time passes after a tab goes backgrounded, the
backgrounded tab's memory cache is purged and suspended. We expect
only purgable memory like cache is purged and tab behaviors would never
be affected.

This CL adds only suspending and not purging any cache. I'll add
purging feature later.

BUG=551995, 607077

Committed: https://crrev.com/6909327ee89d7bd5b69e9e578bcb28388271b2ea
Cr-Commit-Position: refs/heads/master@{#393476}
==========

[commit-bot: I haz the power, 2016-05-13 08:32:51]

Patchset 13 (id:??) landed as
https://crrev.com/6909327ee89d7bd5b69e9e578bcb28388271b2ea
Cr-Commit-Position: refs/heads/master@{#393476}