cc: Enable non-opaque harware overlays.

cc: Enable non-opaque harware overlays.

This CL enables alpha-blended HW overlays.


Additionally, it clears up the way overlay candidates might or might not
be considered based on buffer format and alpha blending requested:

- An opacity on the quad different than 1 will cause the candidate to
  always be rejected since we can't modulate the alpha of an overlay plane.

- We support only kSrc (no blending) and kSrcOver (blend) blending modes.

- For fullscreen overlays we can allow both blending modes as long as the
  opaque rect is as big as the buffer (ShouldDrawWithBlending() true).
  An opaque FD will always be used in this case (crrev.com/2743403005).

- For non-fullscreen overlays, we can't allow kSrc if the buffer has a
  format with an alpha channel, since the alpha channel might not be
  initialized and we don't disable blending when scanning out yet.

BUG=695296
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel

Review-Url: https://codereview.chromium.org/2749783006
Cr-Commit-Position: refs/heads/master@{#457528}
Committed: https://chromium.googlesource.com/chromium/src/+/21fe8132924302ea1d830c3df055725cd9efeb8b

[Daniele Castagna, 2017-03-16 04:40:12]

Description was changed from

==========
cc: Enable non-opaque harware overlays.

This CL enables alpha-blended HW overlays.


Additionally, it clears up the way overlay candidates might or might not
be considered based on buffer format and alpha blending requested:

- An opacity on the quad different than 1 will cause the candidate to
  always be rejected since we can't modulate the alpha of an overlay plane.

- We support only kSrc (no blending) and kSrcOver (blend) blending modes.

- For fullscreen overlays we can allow both blending modes as long as the
  opaque rect is as big as the buffer (ShouldDrawWithBlending() true).
  An opaque FD will always be used in this case (crrev.com/2743403005).

- For non-fullscreen overlays, we can't allow kSrc if the buffer has a
  format with an alpha channel, since the alpha channel might not be
  initialized and we don't disable blending when scanning out yet.

BUG=695296
==========

to

==========
cc: Enable non-opaque harware overlays.

This CL enables alpha-blended HW overlays.


Additionally, it clears up the way overlay candidates might or might not
be considered based on buffer format and alpha blending requested:

- An opacity on the quad different than 1 will cause the candidate to
  always be rejected since we can't modulate the alpha of an overlay plane.

- We support only kSrc (no blending) and kSrcOver (blend) blending modes.

- For fullscreen overlays we can allow both blending modes as long as the
  opaque rect is as big as the buffer (ShouldDrawWithBlending() true).
  An opaque FD will always be used in this case (crrev.com/2743403005).

- For non-fullscreen overlays, we can't allow kSrc if the buffer has a
  format with an alpha channel, since the alpha channel might not be
  initialized and we don't disable blending when scanning out yet.

BUG=695296
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel
==========

[Daniele Castagna, 2017-03-16 04:40:16]

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

[commit-bot: I haz the power, 2017-03-16 04:40:30]

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

[Daniele Castagna, 2017-03-16 04:49:15]

dcastagna@chromium.org changed reviewers:
+ reveman@chromium.org

[Daniele Castagna, 2017-03-16 04:49:16]

I haven't tested this on a real device yet.

If you want to start taking a look and checking that all those alpha checks make
sense.

[reveman, 2017-03-16 04:57:34]

lgtm

https://codereview.chromium.org/2749783006/diff/20001/cc/output/overlay_strat...
File cc/output/overlay_strategy_single_on_top.cc (right):

https://codereview.chromium.org/2749783006/diff/20001/cc/output/overlay_strat...
cc/output/overlay_strategy_single_on_top.cc:14: 
nit: remove blank line

https://codereview.chromium.org/2749783006/diff/20001/cc/output/overlay_strat...
cc/output/overlay_strategy_single_on_top.cc:16: const gfx::BufferFormat
kOverlayFormatsWithAlpha[] = {
nit: add blank line before this

https://codereview.chromium.org/2749783006/diff/20001/cc/output/overlay_strat...
cc/output/overlay_strategy_single_on_top.cc:18: }
nit: add blank line after this :)

[Daniele Castagna, 2017-03-16 05:04:20]

https://codereview.chromium.org/2749783006/diff/20001/cc/output/overlay_strat...
File cc/output/overlay_strategy_single_on_top.cc (right):

https://codereview.chromium.org/2749783006/diff/20001/cc/output/overlay_strat...
cc/output/overlay_strategy_single_on_top.cc:14: 
On 2017/03/16 at 04:57:34, reveman wrote:
> nit: remove blank line

Done.

https://codereview.chromium.org/2749783006/diff/20001/cc/output/overlay_strat...
cc/output/overlay_strategy_single_on_top.cc:16: const gfx::BufferFormat
kOverlayFormatsWithAlpha[] = {
On 2017/03/16 at 04:57:34, reveman wrote:
> nit: add blank line before this

Done.

https://codereview.chromium.org/2749783006/diff/20001/cc/output/overlay_strat...
cc/output/overlay_strategy_single_on_top.cc:18: }
On 2017/03/16 at 04:57:34, reveman wrote:
> nit: add blank line after this :)

Done.

[reveman, 2017-03-16 13:08:38]

FYI, tested this with laser pointer code and it works perfectly

[Daniele Castagna, 2017-03-16 17:52:28]

+halliwell in case this might break something Chromecast related.

[halliwell, 2017-03-16 18:14:27]

On 2017/03/16 17:52:28, Daniele Castagna wrote:
> +halliwell in case this might break something Chromecast related.

lgtm, I will test it tomorrow to be certain, but feel free to land.

[Daniele Castagna, 2017-03-16 18:15:40]

The CQ bit was checked by dcastagna@chromium.org

[Daniele Castagna, 2017-03-16 18:15:41]

The patchset sent to the CQ was uploaded after l-g-t-m from reveman@chromium.org
Link to the patchset: https://codereview.chromium.org/2749783006/#ps40001
(title: "Blank lines!")

[commit-bot: I haz the power, 2017-03-16 18:16:17]

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, 2017-03-16 20:04:28]

CQ is committing da patch.

Bot data: {"patchset_id": 40001, "attempt_start_ts": 1489688140571680,
"parent_rev": "4576d3a030396c7542eeb7e07017e4ac1e9ffa28", "commit_rev":
"21fe8132924302ea1d830c3df055725cd9efeb8b"}

[commit-bot: I haz the power, 2017-03-16 20:05:06]

Description was changed from

==========
cc: Enable non-opaque harware overlays.

This CL enables alpha-blended HW overlays.


Additionally, it clears up the way overlay candidates might or might not
be considered based on buffer format and alpha blending requested:

- An opacity on the quad different than 1 will cause the candidate to
  always be rejected since we can't modulate the alpha of an overlay plane.

- We support only kSrc (no blending) and kSrcOver (blend) blending modes.

- For fullscreen overlays we can allow both blending modes as long as the
  opaque rect is as big as the buffer (ShouldDrawWithBlending() true).
  An opaque FD will always be used in this case (crrev.com/2743403005).

- For non-fullscreen overlays, we can't allow kSrc if the buffer has a
  format with an alpha channel, since the alpha channel might not be
  initialized and we don't disable blending when scanning out yet.

BUG=695296
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel
==========

to

==========
cc: Enable non-opaque harware overlays.

This CL enables alpha-blended HW overlays.


Additionally, it clears up the way overlay candidates might or might not
be considered based on buffer format and alpha blending requested:

- An opacity on the quad different than 1 will cause the candidate to
  always be rejected since we can't modulate the alpha of an overlay plane.

- We support only kSrc (no blending) and kSrcOver (blend) blending modes.

- For fullscreen overlays we can allow both blending modes as long as the
  opaque rect is as big as the buffer (ShouldDrawWithBlending() true).
  An opaque FD will always be used in this case (crrev.com/2743403005).

- For non-fullscreen overlays, we can't allow kSrc if the buffer has a
  format with an alpha channel, since the alpha channel might not be
  initialized and we don't disable blending when scanning out yet.

BUG=695296
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel

Review-Url: https://codereview.chromium.org/2749783006
Cr-Commit-Position: refs/heads/master@{#457528}
Committed:
https://chromium.googlesource.com/chromium/src/+/21fe8132924302ea1d830c3df055...
==========

[commit-bot: I haz the power, 2017-03-16 20:05:07]

Committed patchset #3 (id:40001) as
https://chromium.googlesource.com/chromium/src/+/21fe8132924302ea1d830c3df055...