Move client_native_pixmap_factory.h from ui/ozone/public to ui/gfx

ui/gfx/client_native_pixmap.h

 // Copyright 2015 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
-#ifndef UI_OZONE_PUBLIC_CLIENT_NATIVE_PIXMAP_H_
-#define UI_OZONE_PUBLIC_CLIENT_NATIVE_PIXMAP_H_
+#ifndef UI_GFX_CLIENT_NATIVE_PIXMAP_H_
+#define UI_GFX_CLIENT_NATIVE_PIXMAP_H_
 
-#include "ui/ozone/ozone_export.h"
+#include "ui/gfx/gfx_export.h"
 
 namespace ui {
 
 // This represents a buffer that can be written to directly by regular CPU code,
 // but can also be read by the GPU.
 // NativePixmap is its counterpart in GPU process.
-class OZONE_EXPORT ClientNativePixmap {
+class GFX_EXPORT ClientNativePixmap {

[reveman, 2017/03/08 16:41:55]

Would be nice to rename this and relevant classes to something like DmaBufHolder
in a follow up if this is now specific to dmabufs.

[spang, 2017/03/08 17:43:44]

There's nothing in here that is specific to dmabufs. This is for anything that
supports import and mmap.

[reveman, 2017/03/09 12:10:48]

There's the SYNC_START/END logic.
https://cs.chromium.org/chromium/src/ui/ozone/platform/drm/common/client_nati...

Maybe we don't even need a "holder" class on the client side but can just move
this into the GMB impl. I think it would be nice to have this logic in a
gpu_memory_buffer_impl_linux_dmabuf.cc file and no longer be tied up in ozone
specific code. One layer less of abstraction..

  public:
   virtual ~ClientNativePixmap() {}
 
   // Map each plane in the client address space.
   // Return false on error.
   virtual bool Map() = 0;
   virtual void Unmap() = 0;
 
   virtual void* GetMemoryAddress(size_t plane) const = 0;
   virtual int GetStride(size_t plane) const = 0;
 };
 
 }  // namespace ui
 
-#endif  // UI_OZONE_PUBLIC_CLIENT_NATIVE_PIXMAP_H_
+#endif  // UI_GFX_CLIENT_NATIVE_PIXMAP_H_