Introduce base::MemoryPressureMonitor

Introduce base::MemoryPressureMonitor

This is a step towards unifying the various ways to "notify"
MemoryPressureListeners.

This also provides "GetCurrentPressureLevel" to a more general-purpose
place so we'll be able to use it on other platforms.
(Especially Windows, so we can use it in renderer_frame_manager).

See:
https://docs.google.com/a/chromium.org/document/d/1z3SR3JfKOwVaZxrsfHLOrPjo4-IVtnijaX86LKYPEQs/edit#
for more on the rationale and design, and
https://codereview.chromium.org/1045433002/
for what this might look like for ChromeOS's memory
monitor.

BUG=463546
Committed: https://crrev.com/87f6aa1c21c81727a9d6e9813b66dbc58eb2ecc0
Cr-Commit-Position: refs/heads/master@{#324086}

[dmichael (off chromium), 2015-03-31 20:34:56]

dmichael@chromium.org changed reviewers:
+ chrisha@chromium.org, skuhne@chromium.org

[dmichael (off chromium), 2015-03-31 21:12:25]

split from https://codereview.chromium.org/1045433002/

[dmichael (off chromium), 2015-04-03 23:04:08]

dmichael@chromium.org changed reviewers:
+ brettw@chromium.org, jam@chromium.org

[dmichael (off chromium), 2015-04-03 23:04:09]

jam and brettw, how do you feel about having a GetCurrentPressureLevel available
in base/? It will help us with using some of ChromeOS's work on other platforms.

[jam, 2015-04-03 23:11:31]

On 2015/04/03 23:04:09, dmichael wrote:
> jam and brettw, how do you feel about having a GetCurrentPressureLevel
available
> in base/? It will help us with using some of ChromeOS's work on other
platforms.

I'm not really a base owner, so you should ask the folks in base/OWNERS

[dmichael (off chromium), 2015-04-06 16:40:06]

dmichael@chromium.org changed reviewers:
+ cpu@chromium.org, danakj@chromium.org

[dmichael (off chromium), 2015-04-06 16:40:07]

Thanks, jam@. I added you because brettw indicated you might have strong
opinions around what is exposed for Memory Pressure. He also suggested
chrome-eng-review, but I think what we want is less than what I originally asked
him about (I thought we wanted an event for "no memory pressure", but we don't
need that).

+danakj as an OWNER, and +cpu as somebody else who may have strong opinions
around memory pressure.

I'm proposing adding a way to get the current pressure level. Currently, the
only memory pressure API available in base/ is receiving events for "moderate"
and "critical" memory pressure via MemoryPressureListener.

I'd appreciate any help finding the right stakeholders, so we can try to figure
out the best way to make this info available.

[danakj, 2015-04-06 16:47:21]

danakj@chromium.org changed reviewers:
+ avi@chromium.org

[danakj, 2015-04-06 16:47:21]

+avi

[Avi (use Gerrit), 2015-04-06 17:44:05]

On 2015/04/06 16:47:21, danakj wrote:
> +avi

What do you want my opinion on? This is more of a question of "does this belong
in base/" and I usually get that question about content/.

[danakj, 2015-04-06 17:48:35]

On 2015/04/06 17:44:05, Avi (OOO Friday 3 Apr) wrote:
> On 2015/04/06 16:47:21, danakj wrote:
> > +avi
> 
> What do you want my opinion on? This is more of a question of "does this
belong
> in base/" and I usually get that question about content/.

You reviewed all the browser discardable memory stuff, so I thought you might be
interested in knowing about this, and possibly have opinions on what it should
look like.

[Avi (use Gerrit), 2015-04-06 18:52:43]

On 2015/04/06 17:48:35, danakj wrote:
> On 2015/04/06 17:44:05, Avi (OOO Friday 3 Apr) wrote:
> > On 2015/04/06 16:47:21, danakj wrote:
> > > +avi
> > 
> > What do you want my opinion on? This is more of a question of "does this
> belong
> > in base/" and I usually get that question about content/.
> 
> You reviewed all the browser discardable memory stuff, so I thought you might
be
> interested in knowing about this, and possibly have opinions on what it should
> look like.

The design looks reasonable to me. The code itself is just a shell so I don't
have much to say about it.

[danakj, 2015-04-07 00:13:24]

This seems fine to me.

Did you want to land this as an atomic piece or add to it before landing?

https://codereview.chromium.org/1048983006/diff/20001/base/memory/memory_pres...
File base/memory/memory_pressure_monitor.cc (right):

https://codereview.chromium.org/1048983006/diff/20001/base/memory/memory_pres...
base/memory/memory_pressure_monitor.cc:17: CHECK(!g_monitor);
DCHECK please

https://codereview.chromium.org/1048983006/diff/20001/base/memory/memory_pres...
base/memory/memory_pressure_monitor.cc:22: CHECK(g_monitor);
ditto

[cpu_(ooo_6.6-7.5), 2015-04-07 00:47:11]

ping me if this CL gets meat, specially on windows.

[dmichael (off chromium), 2015-04-07 16:15:57]

danakj: Thanks, review comments addressed.

chrisha asked me to split the earlier version up
(https://codereview.chromium.org/1045433002/), hence this CL. It's unfortunate
it's less obvious that it's useful this way, but hopefully the commit
description and doc help justify it.

[danakj, 2015-04-07 18:06:35]

OK LGTM.  Btw, can you try wrap your patch description at 72 chars?

[dmichael (off chromium), 2015-04-07 18:20:25]

On 2015/04/07 18:06:35, danakj wrote:
> OK LGTM.  Btw, can you try wrap your patch description at 72 chars?

Done (save a URL). Thanks!

[dmichael (off chromium), 2015-04-07 18:20:29]

The CQ bit was checked by dmichael@chromium.org

[commit-bot: I haz the power, 2015-04-07 18:21:11]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1048983006/60001

[commit-bot: I haz the power, 2015-04-07 19:13:22]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2015-04-07 19:14:41]

Patchset 4 (id:??) landed as
https://crrev.com/87f6aa1c21c81727a9d6e9813b66dbc58eb2ecc0
Cr-Commit-Position: refs/heads/master@{#324086}

[Mr4D (OOO till 08-26), 2015-04-13 18:04:18]

Sorry - was OOO - lgtm!