D3D, like most graphics APIs, queues & batches up commands. Many things prevent resource memory, like a texture, from being free'ed during Release. To start, there are batched up commands that reference the memory. Flush() clears out the command queue. After command queue references are gone, the GPU may still be using the memory. So, it ...

ResizeBuffers takes the flags as well. ResizeBuffers is expected to be called by your WM_SIZE handler, which will be invoked after you call SetFullscreenState( TRUE ).

I'm unsure what your 'typical' Update/Render paradigm entails, but the concept is rather simple. The Update thread usually hands off data to the Render thread. If you only have one piece of data, then the data is usually protected with synchronization (which would only let one thread manipulate it at a time); and the Update thread ...

There are multiple forms of multi-buffering. The multi-buffering the swap chain does has more to do with preventing rendering to a front buffer (which is typically slow and can produce tearing). Multi-buffering also is used to enable parallelism between multiple processing units. Traditionally, muti-buffering was used to ensure the GPU would ...

When does Present occur, how many back buffers, and what is the present effect? If Present occurs before CopyResource, then it is risky what you are retrieving.. Present will rotate the back buffers identities, such that back buffer 1 is now 0, etc.

IDXGISwapChain inherits from an interface which has GetDevice on it. You might be able to avoid explicitly hooking the device.

Those fields are used as full-screen display mode preferences. By default, DXGI swap chains hook Alt-Enter to provide applications with the ability to transition from Windowed to Fullscreen. When fullscreen, those fields indicate what type of preferences the application would like. As they are only preferences, numerous other factors affect which ...

Are you processing WM_SIZE? Most applications are expected to process WM_SIZE and call IDXGISwapChain::ResizeBuffers. The swap chain is supposed to continue to function to some degree even if the application doesn't have WM_SIZE processing; but I'm not immediately sure what's wrong with that scenario.

Unfortunately, the debugging infrastructure does not point out leaks about core objects (e.g. the Device). It is an often requested feature. In the mean time, I believe PIX for Windows can point out such leaks.

Nope. If I remember correctly, the reason was because it could be emulated in a shader; and was deemed not popular enough to warrant the additional API special case.