Port SkBlurImage opts to SkOpts.

Port SkBlurImage opts to SkOpts.

+268 -535 lines

I also rearranged the code a little bit to encapsulate itself better,
mostly replacing static helper functions with lambdas.  This also
let me merge the SSE2 and SSE4.1 code paths.

BUG=skia:4117
Committed: https://skia.googlesource.com/skia/+/dce5ce4276e2825efc6d8c4daa819c965794cd12

[mtklein_C, 2015-08-04 00:17:27]

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

[commit-bot: I haz the power, 2015-08-04 00:17:36]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1264103004/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1264103004/1

[commit-bot: I haz the power, 2015-08-04 00:21:27]

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

[commit-bot: I haz the power, 2015-08-04 00:21:28]

Dry run: Try jobs failed on following builders:
  Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Debug-Trybot on client.skia (JOB_FAILED,
http://build.chromium.org/p/client.skia/builders/Test-Ubuntu-GCC-GCE-CPU-AVX2...)

[mtklein, 2015-08-04 00:35:05]

The CQ bit was checked by mtklein@google.com to run a CQ dry run

[commit-bot: I haz the power, 2015-08-04 00:35:20]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1264103004/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1264103004/40001

[commit-bot: I haz the power, 2015-08-04 00:42:01]

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

[commit-bot: I haz the power, 2015-08-04 00:42:01]

Dry run: This issue passed the CQ dry run.

[mtklein_C, 2015-08-04 00:42:08]

mtklein@chromium.org changed reviewers:
+ djsollen@google.com
- mtklein@google.com

[djsollen, 2015-08-04 14:39:57]

https://codereview.chromium.org/1264103004/diff/40001/src/core/SkOpts.cpp
File src/core/SkOpts.cpp (right):

https://codereview.chromium.org/1264103004/diff/40001/src/core/SkOpts.cpp#new...
src/core/SkOpts.cpp:10: #define SK_OPTS_NS portable
given that you don't yet use this define I would remove it from this CL

https://codereview.chromium.org/1264103004/diff/40001/src/core/SkOpts.cpp#new...
src/core/SkOpts.cpp:51: decltype(box_blur_xx) box_blur_xx = nullptr;
this is a break from previous versions where we default to the portable code. 
Why the switch?

https://codereview.chromium.org/1264103004/diff/40001/src/effects/SkBlurImage...
File src/effects/SkBlurImageFilter.cpp (right):

https://codereview.chromium.org/1264103004/diff/40001/src/effects/SkBlurImage...
src/effects/SkBlurImageFilter.cpp:208: auto boxBlurX  = SkOpts::box_blur_xx,
instead of Auto would it be more clear if we used SkOpts::BoxBlur?

https://codereview.chromium.org/1264103004/diff/40001/src/opts/SkXfermode_opts.h
File src/opts/SkXfermode_opts.h (right):

https://codereview.chromium.org/1264103004/diff/40001/src/opts/SkXfermode_opt...
src/opts/SkXfermode_opts.h:15: namespace /* TODO: SK_OPTS_NS */ {
I don't think there is a need to touch this file in this CL.

[mtklein, 2015-08-04 14:46:44]

mtklein@google.com changed reviewers:
+ mtklein@google.com

[mtklein, 2015-08-04 14:46:45]

https://codereview.chromium.org/1264103004/diff/40001/src/core/SkOpts.cpp
File src/core/SkOpts.cpp (right):

https://codereview.chromium.org/1264103004/diff/40001/src/core/SkOpts.cpp#new...
src/core/SkOpts.cpp:10: #define SK_OPTS_NS portable
On 2015/08/04 14:39:56, djsollen wrote:
> given that you don't yet use this define I would remove it from this CL

Let's leave it.  It's what it wants to be, modulo me figuring out why I can't
make it work in SkXfermode_opt.h.

https://codereview.chromium.org/1264103004/diff/40001/src/core/SkOpts.cpp#new...
src/core/SkOpts.cpp:51: decltype(box_blur_xx) box_blur_xx = nullptr;
On 2015/08/04 14:39:56, djsollen wrote:
> this is a break from previous versions where we default to the portable code. 
> Why the switch?

The portable blurs are in src/effects, and this keeps it that way.  We can
follow up by moving them to src/opts if you like.  I can see arguments for both
ways.

https://codereview.chromium.org/1264103004/diff/40001/src/effects/SkBlurImage...
File src/effects/SkBlurImageFilter.cpp (right):

https://codereview.chromium.org/1264103004/diff/40001/src/effects/SkBlurImage...
src/effects/SkBlurImageFilter.cpp:208: auto boxBlurX  = SkOpts::box_blur_xx,
On 2015/08/04 14:39:56, djsollen wrote:
> instead of Auto would it be more clear if we used SkOpts::BoxBlur?

Both seem fine to me.  Do you prefer SkOpts::BoxBlur?

https://codereview.chromium.org/1264103004/diff/40001/src/opts/SkXfermode_opts.h
File src/opts/SkXfermode_opts.h (right):

https://codereview.chromium.org/1264103004/diff/40001/src/opts/SkXfermode_opt...
src/opts/SkXfermode_opts.h:15: namespace /* TODO: SK_OPTS_NS */ {
On 2015/08/04 14:39:56, djsollen wrote:
> I don't think there is a need to touch this file in this CL.

I want to keep everything consistently using SK_OPTS_NS.

However, using SK_OPTS_NS here triggers a warning with some compilers.  I want
to sort that out in another CL.

[djsollen, 2015-08-04 14:56:17]

https://codereview.chromium.org/1264103004/diff/40001/src/core/SkOpts.cpp
File src/core/SkOpts.cpp (right):

https://codereview.chromium.org/1264103004/diff/40001/src/core/SkOpts.cpp#new...
src/core/SkOpts.cpp:51: decltype(box_blur_xx) box_blur_xx = nullptr;
On 2015/08/04 14:46:44, mtklein wrote:
> On 2015/08/04 14:39:56, djsollen wrote:
> > this is a break from previous versions where we default to the portable
code. 
> > Why the switch?
> 
> The portable blurs are in src/effects, and this keeps it that way.  We can
> follow up by moving them to src/opts if you like.  I can see arguments for
both
> ways.

I like the idea that SkOpts functions are always defined and you don't need to
do these null checks, but I'm ok to put that off to a follow up CL.

https://codereview.chromium.org/1264103004/diff/40001/src/effects/SkBlurImage...
File src/effects/SkBlurImageFilter.cpp (right):

https://codereview.chromium.org/1264103004/diff/40001/src/effects/SkBlurImage...
src/effects/SkBlurImageFilter.cpp:208: auto boxBlurX  = SkOpts::box_blur_xx,
On 2015/08/04 14:46:44, mtklein wrote:
> On 2015/08/04 14:39:56, djsollen wrote:
> > instead of Auto would it be more clear if we used SkOpts::BoxBlur?
> 
> Both seem fine to me.  Do you prefer SkOpts::BoxBlur?

Yeah, I find it more clear as a reader.

https://codereview.chromium.org/1264103004/diff/40001/src/opts/SkXfermode_opts.h
File src/opts/SkXfermode_opts.h (right):

https://codereview.chromium.org/1264103004/diff/40001/src/opts/SkXfermode_opt...
src/opts/SkXfermode_opts.h:15: namespace /* TODO: SK_OPTS_NS */ {
On 2015/08/04 14:46:44, mtklein wrote:
> On 2015/08/04 14:39:56, djsollen wrote:
> > I don't think there is a need to touch this file in this CL.
> 
> I want to keep everything consistently using SK_OPTS_NS.
> 
> However, using SK_OPTS_NS here triggers a warning with some compilers.  I want
> to sort that out in another CL.

Acknowledged.

[mtklein_C, 2015-08-04 15:16:56]

https://codereview.chromium.org/1264103004/diff/40001/src/core/SkOpts.cpp
File src/core/SkOpts.cpp (right):

https://codereview.chromium.org/1264103004/diff/40001/src/core/SkOpts.cpp#new...
src/core/SkOpts.cpp:51: decltype(box_blur_xx) box_blur_xx = nullptr;
On 2015/08/04 14:56:17, djsollen wrote:
> On 2015/08/04 14:46:44, mtklein wrote:
> > On 2015/08/04 14:39:56, djsollen wrote:
> > > this is a break from previous versions where we default to the portable
> code. 
> > > Why the switch?
> > 
> > The portable blurs are in src/effects, and this keeps it that way.  We can
> > follow up by moving them to src/opts if you like.  I can see arguments for
> both
> > ways.
> 
> I like the idea that SkOpts functions are always defined and you don't need to
> do these null checks, but I'm ok to put that off to a follow up CL.

I tried it and I'm convinced.  Done!

[mtklein_C, 2015-08-04 15:22:03]

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

[commit-bot: I haz the power, 2015-08-04 15:22:17]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1264103004/80001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1264103004/80001

[djsollen, 2015-08-04 15:36:17]

lgtm

[mtklein_C, 2015-08-04 15:41:34]

The CQ bit was checked by mtklein@chromium.org

[mtklein_C, 2015-08-04 15:41:34]

The CQ bit was unchecked by mtklein@chromium.org

[commit-bot: I haz the power, 2015-08-04 15:41:43]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1264103004/80001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1264103004/80001

[commit-bot: I haz the power, 2015-08-04 15:49:25]

Committed patchset #5 (id:80001) as
https://skia.googlesource.com/skia/+/dce5ce4276e2825efc6d8c4daa819c965794cd12