These can legitimately occur in some circumstances, and shouldn't otherwise
affect these tests.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Shin Megami Tensei III Nocturne HD Remaster depends on the frame latency
semaphore being released several times after it calls
IDXGISwapChain2::SetMaximumFrameLatency(3).
Signed-off-by: Andrew Eikum <aeikum@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Since the factory and adapter interfaces may be wrapped, calling
IDXGIAdapter_EnumOutputs() on a wrapped adapter may in turn return a wrapped
output interface. Things would go poorly from there, so make sure we have a
unwrapped interface here. Alternatively, we could have allowed output
interfaces to be wrapped, but tests show that e.g.
IDXGISwapChain_SetFullscreenState() fails when passed a wrapped output
interface.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Mostly to test that when doing mode restoration, all display outputs get restored to their display
settings in the registry, not just one output gets restored.
Signed-off-by: Zhiyi Zhang <zzhang@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Some games specifically ignore the last adapter from the adapter list,
as it's always WARP on Windows. This makes them fail to detect GPUs
correctly on Wine, as they drop one of the reported adapters.
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Some of the invalid swapchain parameter tests crash on Win10 1909.
And DXGI_USAGE_DISCARD_ON_PRESENT is a DXGI internal flag. It doesn't
always get added if swap effect is DXGI_SWAP_EFFECT_DISCARD.
Signed-off-by: Zhiyi Zhang <zzhang@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
On Windows, a DXGI output may support 1920x1440 and 2560x1080
display modes, both of which have the same pixel count. And when
finding a mode closest to 2559x1079 or 2561x1081, 1920x1440 is returned.
Signed-off-by: Zhiyi Zhang <zzhang@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
First cast to LONG, then to DWORD.
Reported by MinGW GCC 10 with -Woverflow.
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Calling d3d12_swapchain_Present() right after switching to full screen
mode may fail because the swapchain window may be in the animation of
becoming full screen and vkQueuePresentKHR() returns VK_ERROR_OUT_OF_DATE_KHR,
even after d3d12_swapchain_present() tried to recreate the swapchain.
Signed-off-by: Zhiyi Zhang <zzhang@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Some d3d12 games gate their use of sync interval 0 on support for
DXGI_FEATURE_PRESENT_ALLOW_TEARING.
Signed-off-by: Zhiyi Zhang <zzhang@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
DirectDraw should clip the cursor to the primary monitor only when the
cooperative level is DDSCL_EXCLUSIVE. For Direct3D 8, 9 and DXGI,
changing display modes shouldn't clip the cursor as tests showed.
Signed-off-by: Zhiyi Zhang <zzhang@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
DXGI_MODE_SCALING_CENTERED and DXGI_MODE_SCALING_STRETCHED modes may be
supported by drivers. When these flags are specified, IDXGIOutput_FindClosestMatchingMode()
will try to find a mode matching the scaling field first, which may not
be the same as the original mode. So make sure such a mode exists before
finding them.
Signed-off-by: Zhiyi Zhang <zzhang@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Zhiyi Zhang