Modernize atomics in SkTaskGroup's threadpool.

    - Use SkAtomic<int32_t> for pending work count so we're statically
      forced to operate on it with atomic methods.
    - Replacing old methods like sk_atomic_inc/dec gives us finer control
      over which barriers we need for each operation.

No public API changes.
TBR=reed@google.com

BUG=skia:

Review URL: https://codereview.chromium.org/1193493003

1 files changed, 24 insertions, 15 deletions

diff --git a/src/core/SkTaskGroup.cpp b/src/core/SkTaskGroup.cpp
index 345fd2fcb9..17c61af960 100644
--- a/src/core/SkTaskGroup.cpp
+++ b/src/core/SkTaskGroup.cpp
@@ -1,3 +1,10 @@
+/*
+ * Copyright 2014 Google Inc.
+ *
+ * Use of this source code is governed by a BSD-style license that can be
+ * found in the LICENSE file.
+ */
+
 #include "SkTaskGroup.h"
 
 #include "SkCondVar.h"
@@ -23,21 +30,22 @@ namespace {
 
 class ThreadPool : SkNoncopyable {
 public:
-    static void Add(SkRunnable* task, int32_t* pending) {
+    static void Add(SkRunnable* task, SkAtomic<int32_t>* pending) {
         if (!gGlobal) {  // If we have no threads, run synchronously.
             return task->run();
         }
         gGlobal->add(&CallRunnable, task, pending);
     }
 
-    static void Add(void (*fn)(void*), void* arg, int32_t* pending) {
+    static void Add(void (*fn)(void*), void* arg, SkAtomic<int32_t>* pending) {
         if (!gGlobal) {
             return fn(arg);
         }
         gGlobal->add(fn, arg, pending);
     }
 
-    static void Batch(void (*fn)(void*), void* args, int N, size_t stride, int32_t* pending) {
+    static void Batch(void (*fn)(void*), void* args, int N, size_t stride,
+                      SkAtomic<int32_t>* pending) {
         if (!gGlobal) {
             for (int i = 0; i < N; i++) { fn((char*)args + i*stride); }
             return;
@@ -45,12 +53,13 @@ public:
         gGlobal->batch(fn, args, N, stride, pending);
     }
 
-    static void Wait(int32_t* pending) {
+    static void Wait(SkAtomic<int32_t>* pending) {
         if (!gGlobal) {  // If we have no threads, the work must already be done.
-            SkASSERT(*pending == 0);
+            SkASSERT(pending->load(sk_memory_order_relaxed) == 0);
             return;
         }
-        while (sk_acquire_load(pending) > 0) {  // Pairs with sk_atomic_dec here or in Loop.
+        // Acquire pairs with decrement release here or in Loop.
+        while (pending->load(sk_memory_order_acquire) > 0) {
             // Lend a hand until our SkTaskGroup of interest is done.
             Work work;
             {
@@ -65,7 +74,7 @@ public:
             // This Work isn't necessarily part of our SkTaskGroup of interest, but that's fine.
             // We threads gotta stick together.  We're always making forward progress.
             work.fn(work.arg);
-            sk_atomic_dec(work.pending);  // Release pairs with the sk_acquire_load() just above.
+            work.pending->fetch_add(-1, sk_memory_order_release);  // Pairs with load above.
         }
     }
 
@@ -80,9 +89,9 @@ private:
     static void CallRunnable(void* arg) { static_cast<SkRunnable*>(arg)->run(); }
 
     struct Work {
-        void (*fn)(void*);  // A function to call,
-        void* arg;          // its argument,
-        int32_t* pending;   // then sk_atomic_dec(pending) afterwards.
+        void (*fn)(void*);            // A function to call,
+        void* arg;                    // its argument,
+        SkAtomic<int32_t>* pending;   // then decrement pending afterwards.
     };
 
     explicit ThreadPool(int threads) : fDraining(false) {
@@ -109,9 +118,9 @@ private:
         fThreads.deleteAll();
     }
 
-    void add(void (*fn)(void*), void* arg, int32_t* pending) {
+    void add(void (*fn)(void*), void* arg, SkAtomic<int32_t>* pending) {
         Work work = { fn, arg, pending };
-        sk_atomic_inc(pending);  // No barrier needed.
+        pending->fetch_add(+1, sk_memory_order_relaxed);  // No barrier needed.
         {
             AutoLock lock(&fReady);
             fWork.push(work);
@@ -119,8 +128,8 @@ private:
         }
     }
 
-    void batch(void (*fn)(void*), void* arg, int N, size_t stride, int32_t* pending) {
-        sk_atomic_add(pending, N);  // No barrier needed.
+    void batch(void (*fn)(void*), void* arg, int N, size_t stride, SkAtomic<int32_t>* pending) {
+        pending->fetch_add(+N, sk_memory_order_relaxed);  // No barrier needed.
         {
             AutoLock lock(&fReady);
             Work* batch = fWork.append(N);
@@ -147,7 +156,7 @@ private:
                 pool->fWork.pop(&work);
             }
             work.fn(work.arg);
-            sk_atomic_dec(work.pending);  // Release pairs with sk_acquire_load() in Wait().
+            work.pending->fetch_add(-1, sk_memory_order_release);  // Pairs with load in Wait().
         }
     }