Re: [PATCH v3 0/1] MR2602: mfplat: Fix returned buffer length in dxgi_surface_buffer_lock().

[Paul Gofman (＠gofman)]

v3:
 - move the test before 'subresource 1' test. This way it works with WARP driver (looks like subresource 1 test just brings WARP device into broken state). The texture creation still fails with reference device, and now it looks indeed format related. I tried a couple of other chroma formats but they don't seem to work with reference device, so probably the test still has to be skipped in this case.