https://bugs.winehq.org/show_bug.cgi?id=49395
--- Comment #13 from RĂ©mi Bernon rbernon@codeweavers.com --- Created attachment 69096 --> https://bugs.winehq.org/attachment.cgi?id=69096 Fix
I've sent a first series of tests to validate D2D / D3D11 interop, and I've implemented a fix for this regression, although the changes may be a bit too large for the code freeze.
It uses D3DDeviceContextState to enable D3D10Device interface on D3D11 device, with a semi-stub implementation in wined3d to allow interface query again after SwapDeviceContextState has been called. The state is not actually restored, but I don't think Wine D2D implementation depends on it, as it does state record / restore on the D3D10 interface already.
I'm attaching the patch series here, in any case.