Mac Overlays: Add AllOrNothing strategy

Mac Overlays: Add AllOrNothing strategy

This strategy attempts to promote all quads into overlays.
If any quad can't be promoted, no quads are promoted.

BUG=533687
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel

Committed: https://crrev.com/351654665d1cce4d5297e467392c42e136734ab7
Cr-Commit-Position: refs/heads/master@{#351896}

[Andre, 2015-09-29 22:20:23]

Patchset #1 (id:1) has been deleted

[Andre, 2015-09-29 23:07:49]

andresantoso@chromium.org changed reviewers:
+ ccameron@chromium.org

[Andre, 2015-09-29 23:07:50]

ccameron PTAL.

[ccameron, 2015-09-29 23:29:56]

lg -- couple of suggestions.

Let's try to resolve this against https://codereview.chromium.org/1376883004/,
too (I don't think it needs any changes -- the primary plane is only a function
of the things that derive from Common).

https://codereview.chromium.org/1373373003/diff/20001/cc/output/overlay_candi...
File cc/output/overlay_candidate.cc (right):

https://codereview.chromium.org/1373373003/diff/20001/cc/output/overlay_candi...
cc/output/overlay_candidate.cc:61: candidate->display_rect = GetOverlayRect(
Remove GetOverlayRect as a function and make it inline here -- having it next to
quad_rect_in_target_space will drive home the point that they two are the same.

https://codereview.chromium.org/1373373003/diff/20001/cc/output/overlay_candi...
File cc/output/overlay_candidate.h (right):

https://codereview.chromium.org/1373373003/diff/20001/cc/output/overlay_candi...
cc/output/overlay_candidate.h:74: OverlayCandidate* candidate);
I'd suggest making GetOverlayTransform be a non-member function, just inside
namespace {} in the .cc.

https://codereview.chromium.org/1373373003/diff/20001/cc/output/overlay_candi...
cc/output/overlay_candidate.h:77: bool y_flipped);
I'd prefer to name "ModifyTransform" to be "ComposeTransforms" with the argument
order reversed, so ComposeTransforms(a, b)(x) = a(b(x)) -- that would me more in
line with the mathematical meaning of the transform.

https://codereview.chromium.org/1373373003/diff/20001/cc/output/overlay_strat...
File cc/output/overlay_strategy_all_or_nothing.cc (right):

https://codereview.chromium.org/1373373003/diff/20001/cc/output/overlay_strat...
cc/output/overlay_strategy_all_or_nothing.cc:34: it->plane_z_order =
next_z_order++;
This can be rolled into the above loop by just having the z orders be -1, -2,
-3, (we will not schedule plane for z=0 in this strategy).

[Andre, 2015-09-30 00:18:38]

Now resolved with https://codereview.chromium.org/1376883004/.
PTAL.

https://codereview.chromium.org/1373373003/diff/20001/cc/output/overlay_candi...
File cc/output/overlay_candidate.cc (right):

https://codereview.chromium.org/1373373003/diff/20001/cc/output/overlay_candi...
cc/output/overlay_candidate.cc:61: candidate->display_rect = GetOverlayRect(
On 2015/09/29 23:29:55, ccameron wrote:
> Remove GetOverlayRect as a function and make it inline here -- having it next
to
> quad_rect_in_target_space will drive home the point that they two are the
same.

Done.

https://codereview.chromium.org/1373373003/diff/20001/cc/output/overlay_candi...
File cc/output/overlay_candidate.h (right):

https://codereview.chromium.org/1373373003/diff/20001/cc/output/overlay_candi...
cc/output/overlay_candidate.h:74: OverlayCandidate* candidate);
On 2015/09/29 23:29:56, ccameron wrote:
> I'd suggest making GetOverlayTransform be a non-member function, just inside
> namespace {} in the .cc.

Done.

https://codereview.chromium.org/1373373003/diff/20001/cc/output/overlay_candi...
cc/output/overlay_candidate.h:77: bool y_flipped);
On 2015/09/29 23:29:55, ccameron wrote:
> I'd prefer to name "ModifyTransform" to be "ComposeTransforms" with the
argument
> order reversed, so ComposeTransforms(a, b)(x) = a(b(x)) -- that would me more
in
> line with the mathematical meaning of the transform.

Done.

https://codereview.chromium.org/1373373003/diff/20001/cc/output/overlay_strat...
File cc/output/overlay_strategy_all_or_nothing.cc (right):

https://codereview.chromium.org/1373373003/diff/20001/cc/output/overlay_strat...
cc/output/overlay_strategy_all_or_nothing.cc:34: it->plane_z_order =
next_z_order++;
On 2015/09/29 23:29:56, ccameron wrote:
> This can be rolled into the above loop by just having the z orders be -1, -2,
> -3, (we will not schedule plane for z=0 in this strategy).

Done.

[ccameron, 2015-09-30 01:27:36]

lgtm

Go and land this ahead without my CL (it looks like it won't get the last stamp
it needs until tomorrow -- I can resolve against this).

[Andre, 2015-09-30 16:15:10]

On 2015/09/30 01:27:36, ccameron wrote:
> lgtm
> 
> Go and land this ahead without my CL (it looks like it won't get the last
stamp
> it needs until tomorrow -- I can resolve against this).

I'll just wait for your CL to land first, it's not a problem.
Updated the validator to allow negative z-order, PTAL.

[Andre, 2015-10-01 17:37:16]

Patchset #4 (id:80001) has been deleted

[Andre, 2015-10-01 20:30:24]

The CQ bit was checked by andresantoso@chromium.org

[Andre, 2015-10-01 20:30:24]

The patchset sent to the CQ was uploaded after l-g-t-m from
ccameron@chromium.org
Link to the patchset: https://codereview.chromium.org/1373373003/#ps100001
(title: "Rebase off master")

[commit-bot: I haz the power, 2015-10-01 20:30:48]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1373373003/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1373373003/100001

[commit-bot: I haz the power, 2015-10-01 20:48:01]

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

[commit-bot: I haz the power, 2015-10-01 20:48:02]

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

[Andre, 2015-10-01 21:09:49]

The CQ bit was checked by andresantoso@chromium.org

[commit-bot: I haz the power, 2015-10-01 21:10:10]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1373373003/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1373373003/100001

[commit-bot: I haz the power, 2015-10-01 21:50:28]

Committed patchset #4 (id:100001)

[commit-bot: I haz the power, 2015-10-01 21:51:37]

Patchset 4 (id:??) landed as
https://crrev.com/351654665d1cce4d5297e467392c42e136734ab7
Cr-Commit-Position: refs/heads/master@{#351896}