Rework SSE and NEON Color32 algorithms to be more correct and faster.

Rework SSE and NEON Color32 algorithms to be more correct and faster.

This algorithm changes the blend math, guarded by SK_LEGACY_COLOR32_MATH.  The new math is more correct: it's never off by more than 1, and correct in all the interesting 0x00 and 0xFF edge cases, where the old math was never off by more than 2, and not always correct on the edges.

If you look at tests/BlendTest.cpp, the old code was using the `blend_256_plus1_trunc` algorithm, while the new code uses `blend_256_round_alt`.  Neither uses `blend_perfect`, which is about ~35% slower than `blend_256_round_alt`.

This will require an unfathomable number of rebaselines, first to Skia, then to Blink when I remove the guard.

I plan to follow up with some integer SIMD abstractions that can unify these two implementations into a single algorithm.  This was originally what I was working on here, but the correctness gains seem to be quite compelling.  The only places these two algorithms really differ greatly now is the kernel function, and even there they can really both be expressed abstractly as:
  - multiply 8-bits and 8-bits producing 16-bits
  - add 16-bits to 16-bits, returning the top 8 bits.
All the constants are the same, except SSE is a little faster to keep 8 16-bit inverse alphas, NEON's a little faster to keep 8 8-bit inverse alphas.  I may need to take this small speed win back to unify the two.

We should expect a ~25% speedup on Intel (mostly from unrolling to 8 pixels) and a ~20% speedup on ARM (mostly from using vaddhn to add `color`, round, and narrow back down to 8-bit all into one instruction.

(I am probably missing several more related bugs here.)
BUG=skia:3738, skia:420, chromium:111470
Committed: https://skia.googlesource.com/skia/+/afe2ffb8ba5e7362a2ee6f4e1540c9ab22df2c1e

[mtklein, 2015-04-17 14:24:25]

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

[commit-bot: I haz the power, 2015-04-17 14:24:44]

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

[commit-bot: I haz the power, 2015-04-17 14:24:45]

Dry run: COMMIT=false is deprecated please use the dry run button instead.

[mtklein_C, 2015-04-17 14:24:49]

mtklein@chromium.org changed reviewers:
+ fmalita@chromium.org, reed@google.com
- mtklein@google.com

[mtklein, 2015-04-17 14:50:33]

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

[commit-bot: I haz the power, 2015-04-17 14:50:41]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1092433002/80001

[commit-bot: I haz the power, 2015-04-17 14:57:41]

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

[commit-bot: I haz the power, 2015-04-17 14:57:41]

Dry run: This issue passed the CQ dry run.

[reed1, 2015-04-17 15:18:36]

Lets be sure to capture the new "blessed" src-over algorithm somewhere in the
code/documentation, so that readers know that the new code is explicitly using
this new best-practice.

[mtklein, 2015-04-17 15:52:32]

Done.  Updated the vanilla Color32 too, which we'll see on the Xoom.

[mtklein, 2015-04-17 16:17:43]

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

[commit-bot: I haz the power, 2015-04-17 16:18:14]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1092433002/140001

[commit-bot: I haz the power, 2015-04-17 16:24:16]

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

[commit-bot: I haz the power, 2015-04-17 16:24:16]

Dry run: This issue passed the CQ dry run.

[reed1, 2015-04-17 16:30:11]

thanks for the change/dox in the portable code.

Obviously we'll want to centralize (or remove) where we #define the legacy flag

[reed1, 2015-04-17 16:30:52]

nit: for the most part, we're using SK_SUPPORT_LEGACY_... for these flags, to
make it easy to grep for them.

[mtklein, 2015-04-17 16:36:01]

On 2015/04/17 16:30:52, reed1 wrote:
> nit: for the most part, we're using SK_SUPPORT_LEGACY_... for these flags, to
> make it easy to grep for them.

done

[mtklein, 2015-04-17 16:36:06]

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

[commit-bot: I haz the power, 2015-04-17 16:36:18]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1092433002/160001

[commit-bot: I haz the power, 2015-04-17 16:48:09]

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

[commit-bot: I haz the power, 2015-04-17 16:48:10]

Dry run: This issue passed the CQ dry run.

[mtklein, 2015-04-17 18:00:24]

Going to land this to see the overall effect.  The Chrome guard has landed, so
Chrome and its rolls shouldn't be affected (if they are, that's interesting
too).

[mtklein, 2015-04-17 18:00:28]

The CQ bit was checked by mtklein@google.com

[mtklein, 2015-04-17 18:00:29]

lgtm

[commit-bot: I haz the power, 2015-04-17 18:00:38]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1092433002/160001

[commit-bot: I haz the power, 2015-04-17 18:01:02]

Committed patchset #9 (id:160001) as
https://skia.googlesource.com/skia/+/afe2ffb8ba5e7362a2ee6f4e1540c9ab22df2c1e