3-15% speedup to HardLight / Overlay xfermodes.

src/opts/Sk4px_NEON.h

 /*
  * Copyright 2015 Google Inc.
  *
  * Use of this source code is governed by a BSD-style license that can be
  * found in the LICENSE file.
  */
 
 namespace { // See Sk4px.h
 
 inline Sk4px Sk4px::DupPMColor(SkPMColor px) { return Sk16b((uint8x16_t)vdupq_n_u32(px)); }
@@ skipping 22 matching lines @@
 inline Sk4px::Wide Sk4px::widenLo() const {
     return Sk16h(vmovl_u8(vget_low_u8 (this->fVec)),
                  vmovl_u8(vget_high_u8(this->fVec)));
 }
 
 inline Sk4px::Wide Sk4px::widenHi() const {
     return Sk16h(vshll_n_u8(vget_low_u8 (this->fVec), 8),
                  vshll_n_u8(vget_high_u8(this->fVec), 8));
 }
 
+inline Sk4px::Wide Sk4px::widenLoHi() const {
+    auto zipped = vzipq_u8(this->fVec, this->fVec);
+    return Sk16h((uint16x8_t)zipped.val[0],
+                 (uint16x8_t)zipped.val[1]);
+}
+
 inline Sk4px::Wide Sk4px::mulWiden(const Sk16b& other) const {
     return Sk16h(vmull_u8(vget_low_u8 (this->fVec), vget_low_u8 (other.fVec)),
                  vmull_u8(vget_high_u8(this->fVec), vget_high_u8(other.fVec)));
 }
 
 inline Sk4px Sk4px::Wide::addNarrowHi(const Sk16h& other) const {
     const Sk4px::Wide o(other);  // Should be no code, but allows us to access fLo, fHi.
     return Sk16b(vcombine_u8(vaddhn_u16(this->fLo.fVec, o.fLo.fVec),
                              vaddhn_u16(this->fHi.fVec, o.fHi.fVec)));
 }
@@ skipping 25 matching lines @@
     return Sk16b(vandq_u8(this->fVec, (uint8x16_t)vdupq_n_u32(0xFF << SK_A32_SHIFT)));
 }
 
 inline Sk4px Sk4px::zeroAlphas() const {
     // vbic(a,b) == a & ~b
     return Sk16b(vbicq_u8(this->fVec, (uint8x16_t)vdupq_n_u32(0xFF << SK_A32_SHIFT)));
 }
 
 } // namespace