Unify some SkNx code

 - one base case and one N=1 case instead of two each (or three with doubles)
 - use SkNx_cast instead of FromBytes/toBytes
 - 4-at-a-time Sk4f::ToBytes becomes a special standalone Sk4f_ToBytes

If I did everything right, this'll be perf- and pixel- neutral.

https://gold.skia.org/search2?issue=1526523003&unt=true&query=source_type%3Dgm&master=false

BUG=skia:
CQ_EXTRA_TRYBOTS=client.skia:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot

Review URL: https://codereview.chromium.org/1526523003

3 files changed, 8 insertions, 8 deletions

diff --git a/src/effects/SkColorMatrixFilter.cpp b/src/effects/SkColorMatrixFilter.cpp
index 3f0f12f2d8..a7b2b7a2f7 100644
--- a/src/effects/SkColorMatrixFilter.cpp
+++ b/src/effects/SkColorMatrixFilter.cpp
@@ -258,7 +258,7 @@ static Sk4f clamp_0_1(const Sk4f& x) {
 
 static SkPMColor round(const Sk4f& x) {
     SkPMColor c;
-    (x * Sk4f(255) + Sk4f(0.5f)).toBytes((uint8_t*)&c);
+    SkNx_cast<uint8_t>(x * Sk4f(255) + Sk4f(0.5f)).store((uint8_t*)&c);
     return c;
 }
 
@@ -296,7 +296,7 @@ void SkColorMatrixFilter::filterSpan(const SkPMColor src[], int count, SkPMColor
                 continue;
             }
 
-            Sk4f srcf = Sk4f::FromBytes((const uint8_t*)&src_c) * Sk4f(1.0f/255);
+            Sk4f srcf = SkNx_cast<float>(Sk4b::Load((const uint8_t*)&src_c)) * Sk4f(1.0f/255);
 
             if (0xFF != SkGetPackedA32(src_c)) {
                 srcf = unpremul(srcf);
diff --git a/src/effects/gradients/SkLinearGradient.cpp b/src/effects/gradients/SkLinearGradient.cpp
index 53264e4d8c..8760029795 100644
--- a/src/effects/gradients/SkLinearGradient.cpp
+++ b/src/effects/gradients/SkLinearGradient.cpp
@@ -150,7 +150,7 @@ SkLinearGradient::LinearGradientContext::LinearGradientContext(
         const Sk4f scale(1, 1, 1, paintAlpha);
         for (int i = 0; i < count; ++i) {
             uint32_t c = SkSwizzle_Color_to_PMColor(shader.fOrigColors[i]);
-            rec[i].fColor = Sk4f::FromBytes((const uint8_t*)&c) * scale;
+            rec[i].fColor = SkNx_cast<float>(Sk4b::Load((const uint8_t*)&c)) * scale;
             if (i > 0) {
                 SkASSERT(rec[i - 1].fPos <= rec[i].fPos);
             }
@@ -162,7 +162,7 @@ SkLinearGradient::LinearGradientContext::LinearGradientContext(
         for (int i = 0; i < count; ++i) {
             SkPMColor pmc = SkPreMultiplyColor(shader.fOrigColors[i]);
             pmc = SkAlphaMulQ(pmc, alphaScale);
-            rec[i].fColor = Sk4f::FromBytes((const uint8_t*)&pmc);
+            rec[i].fColor = SkNx_cast<float>(Sk4b::Load((const uint8_t*)&pmc));
             if (i > 0) {
                 SkASSERT(rec[i - 1].fPos <= rec[i].fPos);
             }
@@ -699,7 +699,7 @@ find_backward(const SkLinearGradient::LinearGradientContext::Rec rec[], float ti
 
 template <bool apply_alpha> SkPMColor trunc_from_255(const Sk4f& x) {
     SkPMColor c;
-    x.toBytes((uint8_t*)&c);
+    SkNx_cast<uint8_t>(x).store((uint8_t*)&c);
     if (apply_alpha) {
         c = SkPreMultiplyARGB(SkGetPackedA32(c), SkGetPackedR32(c),
                               SkGetPackedG32(c), SkGetPackedB32(c));
@@ -751,7 +751,7 @@ template <bool apply_alpha> void ramp(SkPMColor dstC[], int n, const Sk4f& c, co
     Sk4f cd3 = cd1 + dc2;
     while (n >= 4) {
         if (!apply_alpha) {
-            Sk4f::ToBytes((uint8_t*)dstC, cd0, cd1, cd2, cd3);
+            Sk4f_ToBytes((uint8_t*)dstC, cd0, cd1, cd2, cd3);
             dstC += 4;
         } else {
             *dstC++ = trunc_from_255<apply_alpha>(cd0);
diff --git a/src/effects/gradients/SkRadialGradient.cpp b/src/effects/gradients/SkRadialGradient.cpp
index 52d06392e1..d734aa056a 100644
--- a/src/effects/gradients/SkRadialGradient.cpp
+++ b/src/effects/gradients/SkRadialGradient.cpp
@@ -307,7 +307,7 @@ void shadeSpan_radial_clamp2(SkScalar sfx, SkScalar sdx, SkScalar sfy, SkScalar
             dR = dR + ddR;
 
             uint8_t fi[4];
-            dist.toBytes(fi);
+            SkNx_cast<uint8_t>(dist).store(fi);
 
             for (int i = 0; i < 4; i++) {
                 *dstC++ = cache[toggle + fi[i]];
@@ -319,7 +319,7 @@ void shadeSpan_radial_clamp2(SkScalar sfx, SkScalar sdx, SkScalar sfy, SkScalar
             Sk4f dist = Sk4f::Min(fast_sqrt(R), max);
 
             uint8_t fi[4];
-            dist.toBytes(fi);
+            SkNx_cast<uint8_t>(dist).store(fi);
             for (int i = 0; i < count; i++) {
                 *dstC++ = cache[toggle + fi[i]];
                 toggle = next_dither_toggle(toggle);