Fix up all the easy virtual ... SK_OVERRIDE cases.

dm/DMGpuGMTask.h

 #ifndef DMGpuGMTask_DEFINED
 #define DMGpuGMTask_DEFINED
 
 #include "DMGpuSupport.h"
 #include "DMReporter.h"
 #include "DMTask.h"
 #include "DMTaskRunner.h"
 #include "SkBitmap.h"
 #include "SkString.h"
 #include "SkTemplates.h"
 #include "gm.h"
 
 // This is the main entry point for drawing GMs with the GPU.
 
 namespace DM {
 
 class GpuGMTask : public GpuTask {
 public:
     GpuGMTask(const char* config,
               Reporter*,
               TaskRunner*,
               skiagm::GMRegistry::Factory,
               GrContextFactory::GLContextType,
               GrGLStandard gpuAPI,
               int sampleCount,
               bool useDFText);
 
-    virtual void draw(GrContextFactory*) SK_OVERRIDE;
-    virtual bool shouldSkip() const SK_OVERRIDE;
-    virtual SkString name() const SK_OVERRIDE { return fName; }
+    void draw(GrContextFactory*) SK_OVERRIDE;
+    bool shouldSkip() const SK_OVERRIDE;
+    SkString name() const SK_OVERRIDE { return fName; }
 
 private:
     SkAutoTDelete<skiagm::GM> fGM;
     const SkString fName;
     const GrContextFactory::GLContextType fContextType;
     GrGLStandard fGpuAPI;
     const int fSampleCount;
     const bool fUseDFText;
 };
 
 }  // namespace DM
 
 #endif  // DMGpuGMTask_DEFINED