Commit graph

932 commits

Author SHA1 Message Date
Christoph Haag 351dd9b9de c/util: Implement wait_image in comp_swapchain 2023-06-08 17:01:12 +01:00
Korcan Hussein 17d8e010c8 xrt: App enabled extension state
Adds support for passing through which extensions an OpenXR client
has enabled to server for drivers/compositors to be able to query.
2023-05-25 23:20:38 +01:00
Jakob Bornecrantz 647fd366c5 c/render: Remove compute default_sampler 2023-05-21 12:03:04 +01:00
Jakob Bornecrantz 27ba15cb24 c/util: Remove comp_swapchain_image samplers
No longer used, they where all the same.
2023-05-21 12:03:04 +01:00
Jakob Bornecrantz 1644812dc5 c/main: Remove layer renderer framebuffer sampler 2023-05-21 12:03:04 +01:00
Jakob Bornecrantz 82453bab57 c/main: Use new samplers
The swapchain code was creating a whole bunch of samplers, two for each image.
The thought was that samplers might depend on format, but this was not the
case. So just add a few common ones on the render_resources structs and use
them everywhere. Also fixes the bleeding distortion problem.
2023-05-21 12:03:04 +01:00
Jakob Bornecrantz ee66760b63 c/main: Log assert case 2023-05-21 12:03:04 +01:00
Jakob Bornecrantz 44d2dbc268 c/main: Init gfx & compute render resources earlier 2023-05-21 12:03:04 +01:00
Jakob Bornecrantz 8d738248b0 c/render: Add and use new samplers 2023-05-21 12:03:04 +01:00
Jakob Bornecrantz 023d5bf232 c/main: Support compute queue in mirror code 2023-05-19 23:36:41 +01:00
Jakob Bornecrantz 44c81b6ddb c/main: Refactor out mirror_to_debug_gui code 2023-05-19 23:36:41 +01:00
Jakob Bornecrantz f359826363 c/main: Fix readback timing information 2023-05-19 23:36:41 +01:00
Jakob Bornecrantz 3b5e84fd6f c/shaders: Add blit compute shader 2023-05-19 23:36:41 +01:00
Jakob Bornecrantz b5a40a1b50 a/vk: Make it possible to select vk format in readback pool 2023-05-19 23:18:29 +01:00
bjorn d16576252f c/main: Fix XRT_COMPOSITOR_FORCE_WAYLAND_DIRECT; 2023-05-16 14:40:10 -07:00
samuel degrande 1ad168cfc5 cmake: Fix build failure on msys2/mingw64
This is due to the use of `${SDL2_LIBRARIES}` over `SDL2::SDL2`.

On some 'old' OSes such as Ubuntu 20.04, the SDL2 CMake config does
not set an SDL2:SDL2 target but rather defines SDL2_LIBRARIES and
SDL2_INCLUDE_DIRS variables.

This patch creates an SDL2::SDL2 target, if not already set, based on
those 2 variables.
2023-05-15 16:41:34 +00:00
Jakob Bornecrantz 04d5942fa5 c/main: Try to detect when we miss frames 2023-05-03 12:23:20 +01:00
Jakob Bornecrantz 73f68fd111 c/main: Prefer only two swapchain images 2023-05-03 12:23:19 +01:00
Jakob Bornecrantz 2ab43f7494 c/main: Only do early acquire if we are ready 2023-05-03 12:23:02 +01:00
Jakob Bornecrantz 27fafacf63 c/main: Refactor frame handling a bit 2023-05-03 12:23:02 +01:00
Jakob Bornecrantz 2a17212d7f c/main: Use vk_surface_info when creating images 2023-05-03 12:23:01 +01:00
Fernando Velazquez Innella 6096f2199d c/client: Signal D3D12 fence from GPU instead of CPU 2023-05-02 21:23:19 +00:00
Jakob Bornecrantz cfdea265f4 c/client: Avoid using shared context on synchronization 2023-05-01 14:03:01 +01:00
Jarvis Huang 0d31791092 comp/android: Refine surface creation flow
Reuse MonadoView when "Display over other apps" is enabled. Move surface
creation logic to compositor for consistency. With this approach, compositor
implementer controls the way surface is created.
2023-04-18 16:24:35 -05:00
Jarvis Huang cf7e44b710 aux/android: Refactor MonadoView 2023-04-18 16:22:42 -05:00
Jakob Bornecrantz 7147e47508 c/client: Implement excplicit transitions in Vulkan and D3D12 2023-04-18 14:31:55 +01:00
Jakob Bornecrantz 505e4e8c2f c/client: Refactor image barrier submission into a helper function 2023-04-18 14:21:10 +01:00
Jakob Bornecrantz 0d21b73b05 c/client: Implement transition image stubs 2023-04-18 14:21:10 +01:00
Mario Kleiner ff9ad4016f c/multi: Refine some logging messages a bit.
Add some timestamps and extra logging to the frame scheduling code.

Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
2023-04-15 00:43:20 +00:00
Mario Kleiner ae2e7405dd c/multi: Move log_frame_time_diff() to avoid false late frame reports.
In its original call location, this diagnostic/warning function gets
called for each composition cycle even for client frames which have
not yet been delivered for display, because the frames target display
time "XrFrameEndInfo frameEndInfo.displayTime" (as provided by the
OpenXR client) has not been reached yet. Iow. if a OpenXR client
specifies a target frameEndInfo.displayTime in the future, to request
frame presentation in the future, this will cause a flood of false
"Frame late ..." messages by the compositor, despite nothing being
wrong with the timing, until the frame is actually delivered.

E.g., if frameEndInfo.displayTime is 1 second in the future, we'll
get this for each client xrEndFrame() invocation:

WARN [log_frame_time_diff] Frame late by 11.11ms!
WARN [log_frame_time_diff] Frame late by 22.22ms!
... another 87 like these ...
WARN [log_frame_time_diff] Frame late by 988.43ms!

I think what we want is to only check client frames that are actually
delivered the first time by multi_compositor_deliver_any_frames() for
initial display in the current compositor work cycle, and then report
if this first frame display onset was not on or close to the OpenXR
client requested frameEndInfo.displayTime, but too early or too late,
in violation of the clients wishes.

Moving the call check and call of log_frame_time_diff() achieves this
and gives meaningful debug output.

Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
2023-04-15 00:43:20 +00:00
Jakob Bornecrantz 86262e8b4e a/vk: Remove global command pool 2023-04-14 02:34:49 +01:00
Jakob Bornecrantz e9475b1313 c/main: Use cmd helpers for main submission 2023-04-14 02:34:49 +01:00
Jakob Bornecrantz 231a743896 c/main: Use cmd pool helpers in peek code 2023-04-14 02:34:49 +01:00
Jakob Bornecrantz 185bc2bc3c c/main: Use cmd pool helpers in mirror window code 2023-04-14 02:34:49 +01:00
Jakob Bornecrantz 5ef2b647f1 c/main: Use cmd pool helpers in layer render 2023-04-14 02:34:49 +01:00
Jakob Bornecrantz 1ddfcaa5ab c/client: Use cmd pool helpers 2023-04-14 02:34:49 +01:00
Jakob Bornecrantz d1906d1cf8 c/util: Use cmd pool helpers in swapchain code 2023-04-14 02:34:49 +01:00
Jakob Bornecrantz 197461b4cd c/render: Refactor close loops [NFC] 2023-04-14 02:34:49 +01:00
Jakob Bornecrantz 6598760522 c/render: Put distortion image functions in own file 2023-04-14 02:34:49 +01:00
Jakob Bornecrantz f536ec2009 c/render: Use cmd pool helpers in init code 2023-04-14 02:34:49 +01:00
Jakob Bornecrantz e5fed9c3be c/render: Improve error handling 2023-04-14 02:34:49 +01:00
Jakob Bornecrantz ee0dad22d5 c/main: Tidy readback code [NFC] 2023-04-12 12:49:24 +01:00
Moshi Turner 19e0af55d2 c/main: Bounce readback images off temporary image
Most NVidia graphics cards don't support blitting straight from device memory to host memory, but all GPUs support doing it this way.
2023-04-11 22:27:36 -05:00
Jakob Bornecrantz 9f06b1a25e c/util: Always request sampled in extra bits 2023-04-07 21:44:22 +01:00
Jakob Bornecrantz a03d1421b8 c/client: Respect native compositor's extra usage bits 2023-04-07 21:44:22 +01:00
Jakob Bornecrantz 1adf46b02a c/client: Use correct format in get_swapchain_create_properties funtions 2023-04-07 21:44:20 +01:00
Jakob Bornecrantz d40805e777 c/client: Silence VK_FORMAT_R32_SFLOAT warning in OpenGL code 2023-04-07 14:27:08 +00:00
utzcoz 990ce39fa1 c/main: Remove unused headers for comp_renderer.c 2023-04-07 11:23:43 +00:00
Jakob Bornecrantz 85886e53e1 c/client: Make it clear where we break the OpenXR spec 2023-04-04 13:56:55 +01:00
Jakob Bornecrantz a6764b751a c/client: Don't use vkDeviceWaitIdle
As per the documentation for vkDeviceWaitIdle, it requires that the VkQueue
host access is externally synchronized. And the application may have other
VkQueue's in use that we do not know about.

https://registry.khronos.org/vulkan/specs/1.3-extensions/man/html/vkDeviceWaitIdle.html
2023-04-04 13:56:55 +01:00