I was following the nvprosamples creating the direct display to prevent the OS to affect the usage of it as described in this nvidia sample: [ddisplay sample][1] [1]: https://github.com/nvpro-samples/gl_render_vk_ddisplay
This was approx one year ago, using Vulkan 1.1, everything was working fine.
After upgrade to 1.3 creation of the swapchain was not possible as it seemed that display was not made available for vulkan api anymore, _device->createSwapchainKHRUnique resulting to eErrorInitializationFailed.
This init error was possible to fix by adding the vkAcquireWinrtDisplayNV call which seems that in new version is somehow making the display available for swapchain (surface), same stuff was added to the new version of the nvidia sample.
The problem is that it seems that after the application finishes the monitor is still hanging in some vacuum, invisible for the OS anymore, the process of the application becomes hanging in the taskmanager, without posibility to kill it anyhow (niether admin forced task kill will work as it will state that no instance of the process is running). Even the computer is not possible to be restarted (as the restart screen hangs indefinitely). The Nvidia's sample is causing the same behaviour as mine app.
Is there any way to return the display back to the OS when application ends withou powering off and on the workstation?
I tried to use vkReleaseDisplayEXT on destruction, but without any success (it is also not used anywhere in the sample code if I am not blind).