DM: make GPU tasks multithreaded again.  Big refactor.

The main meat of things is in SkThreadPool.  We can now give SkThreadPool a
type for each thread to create and destroy on its local stack.  It's TLS
without going through SkTLS.

I've split the DM tasks into CpuTasks that run on threads with no TLS, and
GpuTasks that run on threads with a thread local GrContextFactory.

The old CpuTask and GpuTask have been renamed to CpuGMTask and GpuGMTask.

Upshot: default run of out/Debug/dm goes from ~45 seconds to ~20 seconds.

BUG=skia:
R=bsalomon@google.com, mtklein@google.com, reed@google.com

Author: mtklein@chromium.org

Review URL: https://codereview.chromium.org/179233005

git-svn-id: http://skia.googlecode.com/svn/trunk@13632 2bbb7eff-a529-9590-31e7-b0007b416f81
diff --git a/dm/DMTestTask.cpp b/dm/DMTestTask.cpp
index 32a698c..6c3fced 100644
--- a/dm/DMTestTask.cpp
+++ b/dm/DMTestTask.cpp
@@ -8,23 +8,32 @@
 
 namespace DM {
 
+bool TestReporter::allowExtendedTest() const { return FLAGS_pathOpsExtended; }
+bool TestReporter::allowThreaded()     const { return !FLAGS_pathOpsSingleThread; }
+bool TestReporter::verbose()           const { return FLAGS_pathOpsVerbose; }
+
 static SkString test_name(const char* name) {
     SkString result("test ");
     result.append(name);
     return result;
 }
 
-TestTask::TestTask(Reporter* reporter,
-                   TaskRunner* taskRunner,
-                   skiatest::TestRegistry::Factory factory)
-    : Task(reporter, taskRunner)
+CpuTestTask::CpuTestTask(Reporter* reporter,
+                         TaskRunner* taskRunner,
+                         skiatest::TestRegistry::Factory factory)
+    : CpuTask(reporter, taskRunner)
     , fTest(factory(NULL))
     , fName(test_name(fTest->getName())) {}
 
-void TestTask::draw() {
-    if (this->usesGpu()) {
-        fTest->setGrContextFactory(this->getGrContextFactory());
-    }
+GpuTestTask::GpuTestTask(Reporter* reporter,
+                         TaskRunner* taskRunner,
+                         skiatest::TestRegistry::Factory factory)
+    : GpuTask(reporter, taskRunner)
+    , fTest(factory(NULL))
+    , fName(test_name(fTest->getName())) {}
+
+
+void CpuTestTask::draw() {
     fTest->setReporter(&fTestReporter);
     fTest->run();
     if (!fTest->passed()) {
@@ -32,8 +41,13 @@
     }
 }
 
-bool TestTask::TestReporter::allowExtendedTest() const { return FLAGS_pathOpsExtended; }
-bool TestTask::TestReporter::allowThreaded()     const { return !FLAGS_pathOpsSingleThread; }
-bool TestTask::TestReporter::verbose()           const { return FLAGS_pathOpsVerbose; }
+void GpuTestTask::draw(GrContextFactory* grFactory) {
+    fTest->setGrContextFactory(grFactory);
+    fTest->setReporter(&fTestReporter);
+    fTest->run();
+    if (!fTest->passed()) {
+        this->fail(fTestReporter.failure());
+    }
+}
 
 }  // namespace DM