SkRasterPipeline: 8x pipelines, without any 8x code enabled.

Original review here: https://skia-review.googlesource.com/c/2990/
Second attempt here:  https://skia-review.googlesource.com/c/3064/

This is the same as the second attempt, but with the change to SkOpts_hsw.cpp left out.
That omitted part is the key piece... this just lands the refactoring.

CQ_INCLUDE_TRYBOTS=master.client.skia:Perf-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Debug-ASAN-Trybot,Perf-Ubuntu-Clang-GCE-CPU-AVX2-x86_64-Debug-GN,Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD-Trybot,Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-Fast-Trybot;master.client.skia.compile:Build-Win-MSVC-x86_64-Debug-Trybot

GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=3242

Change-Id: Iaafa793a4854c2c9cd7e85cca3701bf871253f71
Reviewed-on: https://skia-review.googlesource.com/3242
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>

1 files changed, 11 insertions, 23 deletions

diff --git a/src/core/SkRasterPipeline.cpp b/src/core/SkRasterPipeline.cpp
index 72d5b7b963..bc7feaccc7 100644
--- a/src/core/SkRasterPipeline.cpp
+++ b/src/core/SkRasterPipeline.cpp
@@ -8,11 +8,12 @@
 #include "SkOpts.h"
 #include "SkRasterPipeline.h"
 
-SkRasterPipeline::SkRasterPipeline() {}
+SkRasterPipeline::SkRasterPipeline() {
+    fBodyStart = SkOpts::body[just_return];
+    fTailStart = SkOpts::tail[just_return];
+}
 
-void SkRasterPipeline::append(SkRasterPipeline::Fn body,
-                              SkRasterPipeline::Fn tail,
-                              void* ctx) {
+void SkRasterPipeline::append(void (*body)(), void (*tail)(), void* ctx) {
     // Each stage holds its own context and the next function to call.
     // So the pipeline itself has to hold onto the first function that starts the pipeline.
     (fBody.empty() ? fBodyStart : fBody.back().fNext) = body;
@@ -20,19 +21,19 @@ void SkRasterPipeline::append(SkRasterPipeline::Fn body,
 
     // Each last stage starts with its next function set to JustReturn as a safety net.
     // It'll be overwritten by the next call to append().
-    fBody.push_back({ &JustReturn, ctx });
-    fTail.push_back({ &JustReturn, ctx });
+    fBody.push_back({ SkOpts::body[just_return], ctx });
+    fTail.push_back({ SkOpts::tail[just_return], ctx });
 }
 
 void SkRasterPipeline::append(StockStage stage, void* ctx) {
-    this->append(SkOpts::stages_4[stage], SkOpts::stages_1_3[stage], ctx);
+    this->append(SkOpts::body[stage], SkOpts::tail[stage], ctx);
 }
 
 void SkRasterPipeline::extend(const SkRasterPipeline& src) {
     SkASSERT(src.fBody.count() == src.fTail.count());
 
-    Fn body = src.fBodyStart,
-       tail = src.fTailStart;
+    auto body = src.fBodyStart,
+         tail = src.fTailStart;
     for (int i = 0; i < src.fBody.count(); i++) {
         SkASSERT(src.fBody[i].fCtx == src.fTail[i].fCtx);
         this->append(body, tail, src.fBody[i].fCtx);
@@ -42,18 +43,5 @@ void SkRasterPipeline::extend(const SkRasterPipeline& src) {
 }
 
 void SkRasterPipeline::run(size_t x, size_t n) {
-    // It's fastest to start uninitialized if the compilers all let us.  If not, next fastest is 0.
-    Sk4f v;
-
-    while (n >= 4) {
-        fBodyStart(fBody.begin(), x,0, v,v,v,v, v,v,v,v);
-        x += 4;
-        n -= 4;
-    }
-    if (n > 0) {
-        fTailStart(fTail.begin(), x,n, v,v,v,v, v,v,v,v);
-    }
+    SkOpts::run_pipeline(x,n, fBodyStart,fBody.begin(), fTailStart,fTail.begin());
 }
-
-void SK_VECTORCALL SkRasterPipeline::JustReturn(Stage*, size_t, size_t, Sk4f,Sk4f,Sk4f,Sk4f,
-                                                                        Sk4f,Sk4f,Sk4f,Sk4f) {}