[WIP] cc: Use reverse spiral iterator in tiling eviction.

[WIP] cc: Use reverse spiral iterator in tiling eviction.

[USE eero AT chromium.org, 2014-10-28 19:20:49]

This CL consists of three changes:
 1) Not to ignore layers without valid priorities during eviction.
    This is achieved by moving eviction priority validity checking from
    layer tree host impl to tiling eviction tile iterators.
 2) https://codereview.chromium.org/644313002/
 3) To return shared tiles only from one tiling eviction tile iterator and
    to restore handling required for activation tiles.

[USE eero AT chromium.org, 2014-10-28 19:21:22]

e.hakkinen@samsung.com changed reviewers:
+ vmpstr@chromium.org

[vmpstr, 2014-10-29 18:26:37]

On 2014/10/28 19:20:49, e_hakkinen wrote:
> This CL consists of three changes:
>  1) Not to ignore layers without valid priorities during eviction.
>     This is achieved by moving eviction priority validity checking from
>     layer tree host impl to tiling eviction tile iterators.
>  2) https://codereview.chromium.org/644313002/
>  3) To return shared tiles only from one tiling eviction tile iterator and
>     to restore handling required for activation tiles.

This is great. Can we split this up into separate patches, so that the review
can be easier? I have some concerns about the way you've implemented (3), since
I don't like that we're passing a lot of knowledge down to the tiling... The
rest of the parts of this can go in sooner though.

[USE eero AT chromium.org, 2014-10-30 18:03:27]

On 2014/10/29 18:26:37, vmpstr wrote:
> Can we split this up into separate patches, so that the review can be easier?

Yes, the first part is in https://codereview.chromium.org/690063002/
The other parts pretty much depend on that.

> I have some concerns about the way you've implemented (3), since
> I don't like that we're passing a lot of knowledge down to the tiling...

At least, it is easy to drop layer_has_twin_layer parameters now that
https://codereview.chromium.org/676953003 has landed.

[USE eero AT chromium.org, 2014-11-14 18:27:00]

Here is a rebase, but I want to go through this once more before review actions
as the first patch set split off from this one transformed quite a much during
review.

I think it is still possible to split this into couple of more smaller patch
sets for easier review.

[USE eero AT chromium.org, 2014-12-12 15:03:33]

Mostly merged (by me and vmpstr), but heavily refactored.