Christoph Haag
2e7451ca5e
comp: Support only mesh and no istortion, remove pano and vive shaders
...
Also changes hardcoded binding indices into variables to make following the code easier
e.g. when adding new UBOs.
2020-09-30 02:05:04 +02:00
Christoph Haag
80f6fc9039
comp: Replace hardcoded binding indices with variables
...
Makes it easier to follow when changing code.
2020-09-30 01:30:30 +02:00
Christoph Haag
fb71c71a8c
xrt: Replace mesh generator with xdev->compute_distortion()
...
Each HMD driver now has to implement compute_distortion() which will be called
by the compositor implementation to generate a mesh (usually).
u_distortion_mesh contains implementations for the defaults (panotools, OpenHMD, vive).
Also adds compute_distortion function for Vive distortion
There are differences between OpenHMD and Panotools values, main differences for now:
* psvr has 5 pano coefficients, ohmd has 3
* psvr uses viewport size and lens center in pixels for distortion calculation, ohmd in meter
* psvr uses different distortion scaling than ohmd
2020-09-28 17:04:15 +01:00
Christoph Haag
e5683d4fde
comp: remove useless pano ubo from mesh distortion
...
It was left in as a debug measure, but is more confusing than useful,
especially with northstar directly generating a mesh and vive with its own shader.
2020-09-22 02:19:13 +02:00
Ryan Pavlik
dae3ab3a4f
xrt: Be sure to include config headers
2020-09-18 13:56:24 +01:00
Jakob Bornecrantz
8e24dda794
xrt: Remove out_timestamp argument to xrt_device::get_tracked_pose
2020-09-07 23:37:09 +01:00
Jakob Bornecrantz
855b2005f3
c/main: Use new xrt_space_graph
2020-09-07 23:33:28 +01:00
Ryan Pavlik
cb26f7255c
comp: Use the correct calloc macro
2020-08-26 11:25:43 -05:00
Jakob Bornecrantz
516212fef4
c/main: Test if a format is supported before adding to list
2020-08-26 14:28:10 +00:00
Jakob Bornecrantz
1eaf4baf91
c/main: Support 16 layers
2020-08-26 15:25:00 +01:00
Jakob Bornecrantz
1496262564
xrt: Constify compositor interface
2020-08-26 15:25:00 +01:00
Christoph Haag
8377655128
comp: Handle skipping rendering of infinite cylinder properly
...
* Setting visibility to none will make the compositor skip rendering this layer
* But do fill in some information in like layer type to ease debugging
2020-08-26 15:59:54 +02:00
Christoph Haag
aecfdfc9f9
comp: Fix typo in switch statement
2020-08-26 15:59:54 +02:00
Jakob Bornecrantz
16b9dc13e2
xrt: Correct assert for ipc data type
2020-08-25 14:55:43 +01:00
Jakob Bornecrantz
cd6a237b4a
c/client: Add support for depth projection
2020-08-25 14:55:43 +01:00
Jakob Bornecrantz
8356308bf9
c/main: Add support for depth projection
2020-08-25 14:55:43 +01:00
Christoph Haag
a9193b0e89
st/oxr: Pass through depth layer to compositor
...
The compositor does not make use of this layer yet.
2020-08-24 10:30:23 +02:00
Christoph Haag
092dddc2da
comp: Support Vk and GL depth and depth stencil formats
...
For now add only the depth formats mandated by OpenGL to maximize the
chances of the Vulkan driver supporting a reasonable set of usage flags
for the formats.
2020-08-24 10:18:12 +02:00
Jakob Bornecrantz
fba9f2532c
c/client: Fix warning and tidy
2020-08-21 18:17:12 +01:00
Christoph Haag
094a9008c9
comp: Add cylinder layer implementation
2020-08-21 02:24:48 +02:00
Christoph Haag
370f6820aa
m: Generalize math_matrix_4x4_quad_model, making scaling a vec3
2020-08-21 02:11:03 +02:00
Christoph Haag
3122e2d073
comp: Use define for number of plane vertices
2020-08-21 02:11:03 +02:00
Christoph Haag
fbe390813c
vk: Add update_buffer helper function
2020-08-21 02:11:03 +02:00
Christoph Haag
cf5596adba
comp: Use CLAMP_TO_EDGE for swapchains
...
Avoids sampling black at the edges of layers.
2020-08-21 02:09:20 +02:00
Ryan Pavlik
640f24438d
comp/client: Clean up commented-out code.
2020-08-19 16:06:30 -05:00
Ryan Pavlik
c49ccd7b46
comp/client: Improve EGLimage-based swapchain
2020-08-19 16:06:30 -05:00
Ryan Pavlik
3250617dcb
comp/client: Use new texture target utility for OpenGL
2020-08-19 16:06:30 -05:00
Ryan Pavlik
521de8ecfd
comp: Add a URL for sample Android EGL code.
2020-08-19 11:01:49 -05:00
Ryan Pavlik
6820a7851d
comp/main: Refuse to create a swapchain if protected content bit requested.
2020-08-14 19:58:01 +00:00
Jakob Bornecrantz
4b5d25dc33
c/main: Stub in code for cube, cylinder and equirect layers
2020-08-14 19:15:02 +00:00
Jakob Bornecrantz
1292173b5e
c/client: Add support for cube, cylinder and equirect layers
2020-08-14 19:15:02 +00:00
Ryan Pavlik
bf37b9d999
comp: Use u_logging in some EGL code.
2020-08-13 17:34:05 -05:00
Ryan Pavlik
1bd2c66aa2
comp: Port to using u_handles
2020-08-13 17:34:05 -05:00
Ryan Pavlik
ab4ceee720
comp: Save/restore the correct texture binding on Android
2020-08-13 17:34:05 -05:00
Ryan Pavlik
f9a613a349
comp: Fix compile-time conditionals
2020-08-13 17:34:05 -05:00
Ryan Pavlik
c5d1dec60c
comp: Move GLAD loading and extension checking to places with more information.
...
They will be sure to load the right combination.
2020-08-13 17:34:05 -05:00
Ryan Pavlik
9087d1c7a9
xrt: Use our generic graphics buffer typedefs all over.
...
This adds some Android support in composition clients,
and fixes the breakage from 2 commits ago.
Thanks to Jakob for finding my error in an earlier version.
2020-08-13 17:34:05 -05:00
Ryan Pavlik
fc73ba0401
comp: Split OpenGL[ES] swapchain types
2020-08-13 17:34:05 -05:00
Ryan Pavlik
475192ec73
xrt: Rename "fd" references to "handle"
...
This breaks some Android pieces, which will be fixed shortly.
2020-08-12 17:11:42 -05:00
Ryan Pavlik
146218b346
xrt: Some clang tidy fixes.
...
A few suppressions as well where clang-tidy did the wrong thing.
2020-08-12 17:11:42 -05:00
Jakob Bornecrantz
71e10fafcd
c/main: Fix warnings with asserts
2020-08-11 01:48:31 +01:00
Christoph Haag
026cc64681
comp: make nvidia display selectable with XRT_COMPOSITOR_FORCE_NVIDIA_DISPLAY
...
Allows temporarily adding a nvidia display name string to the whitelist without recompiling.
2020-08-11 00:44:31 +02:00
Christoph Haag
5ecd0a947d
comp: Move nvidia whitelist warning so it is displayed again
2020-08-10 20:48:29 +02:00
Jakob Bornecrantz
ca786df18c
xrt: Make xrt_swapchain_create_info argument const
2020-08-07 19:15:18 +01:00
Christoph Haag
e48c748a57
xrt: suggest correct GPU in xrGetVulkanGraphicsDevice()
...
As before, on the service side the GPU index the compositor runs on can be selected with
* XRT_COMPOSITOR_FORCE_GPU_INDEX=INDEX1
By default xrGetVulkanGraphicsDevice() will suggest the same GPU the compositor runs on.
It is also possible to override the GPU index suggested to applications with
* XRT_COMPOSITOR_FORCE_CLIENT_GPU_INDEX=INDEX2
The reason this is both done on the service side is that if compositor and client run
on different GPUs, the swapchains use linear tiling instead of optimal tiling.
To make chosen GPUs comparable across the compositor's and the client's vulkan instance,
VkPhysicalDeviceIDProperties.deviceUUID is used.
2020-08-07 17:01:13 +02:00
Jakob Bornecrantz
adc6ca8e8a
c/main: Fill out all of the compositor info struct
2020-08-06 15:14:49 +00:00
Jakob Bornecrantz
ee48b44087
c/main: Remove internal compositor global flip_y
2020-08-06 15:14:49 +00:00
Jakob Bornecrantz
d1d841e7c3
xrt: Use new info struct for compositor formats
2020-08-06 15:14:49 +00:00
Jakob Bornecrantz
e8ca242967
xrt: Remove flip_y argument to native compositor creation
2020-08-06 15:14:49 +00:00
Ryan Pavlik
d8ad620bd2
xrt: Update GLAD again, to a "merged" distribution.
...
This merges the OpenGL and OpenGL-ES code.
2020-08-05 20:50:00 +00:00
