Fixes validation warning when quitting Vulkan OpenXR applications
like hello_xr -G Vulkan2:
VUID-vkDestroyImage-image-01000(ERROR / SPEC): msgNum: -221078694 - Validation Error: [ VUID-vkDestroyImage-image-01000 ] Object 0: handle = 0x55d4d96e4960, type = VK_OBJECT_TYPE_DEVICE; | MessageID = 0xf2d29b5a | Cannot call vkDestroyImage on VkImage 0xcb1c7c000000001b[] that is currently in use by a command buffer. The Vulkan spec states: All submitted commands that refer to image, either directly or via a VkImageView, must have completed execution (https://www.khronos.org/registry/vulkan/specs/1.2-extensions/html/vkspec.html#VUID-vkDestroyImage-image-01000)
Objects: 1
[0] 0x55d4d96e4960, type: 3, name: NULL
This is an addition to commit 778515739f that
introduced this fence.
On recent libsurvive versions, loading the config of the HMD takes quite
long when only a HMD is active without controllers.
Instead of adjusting the timeout value all the time, rework how we wait for
the device configs.
Previously we just waited for device configs to arrive until no new configs
arrive for a predefined timeout.
Now we count how many objects of interest libsurvive actually knows, and then
only wait until the configs for this many devices are loaded.
This should make initialization of the survive driver faster and more robust.
A squash of the following commits.
d/euroc: Add Euroc driver initial boilerplate
d/euroc: Implement xrt_frame_node interface
d/euroc: Implement mainloop thread start flow
d/euroc: Parse samples from data.csv files
d/euroc: Determine dataset information at start
Also use that information to provide a xrt_fs mode and to know if
right camera images are available to use.
d/euroc: Produce frames for left and right sinks
d/euroc: Implement xrt_imu_sink
d/euroc: Implement playback properties
d/euroc: Implement pausing with UI state machine
d/euroc: Show IMU and progress data in UI
d/euroc: Make sure to give in nullptr or valid pointers to wrapL8
d/euroc: Unreference frames when done with them
d/euroc: Fix leaks with debug sinks, and work around free issue
d/euroc: Refactor playback to produce samples with current timestamps
d/euroc: Fix double free by separating debug sinks from downstream sinks
On tegra, when VK_IMAGE_USAGE_COLOR_ATTACHMENT_BIT is set for color image usage
with VK_FORMAT_R16G16B16A16_SFLOAT, dedicated allocation is preferred, but not required.
When using dedicated allocation in this case, we get only a black texture,
and when blitting from that texture, we get a kernel driver error:
gk20a_fifo_handle_mmu_fault_locked:1721 [ERR] fake mmu fault on engine 0, engine subid 1 (hub), client 1 (ce0), addr 0x46cd8aa000, type 15 (poisoned), access_type 0x00000000,inst_ptr 0x212ef3000
Only using dedicated allocation when it is required works around this issue.
Fixes validation warning when acquiring images before the command buffer
of the previous acquire or release on the same queue has finished.
VUID-vkQueueSubmit-pCommandBuffers-00071(ERROR / SPEC): msgNum: 774851941 - Validation Error: [ VUID-vkQueueSubmit-pCommandBuffers-00071 ] Object 0: handle = 0x558634c5c750, type = VK_OBJECT_TYPE_DEVICE; | MessageID = 0x2e2f4d65 | vkQueueSubmit(): pSubmits[0].pCommandBuffers[0] VkCommandBuffer 0x558634b85a10[] is already in use and is not marked for simultaneous use. The Vulkan spec states: If any element of the pCommandBuffers member of any element of pSubmits was not recorded with the VK_COMMAND_BUFFER_USAGE_SIMULTANEOUS_USE_BIT, it must not be in the pending state (https://www.khronos.org/registry/vulkan/specs/1.2-extensions/html/vkspec.html#VUID-vkQueueSubmit-pCommandBuffers-00071)
Objects: 1
[0] 0x558634c5c750, type: 3, name: NULL
qwerty is auto probed, making HMDs that are not auto probed always take precedence.
When setting QWERTY_ENABLE=1 the intent is usually to exclusively use qwerty.
Therefore we default to disabling all other drivers when this variable is set.
To make the old behavior of adding qwerty devices with lower priority than actual
devices, the variable QWERTY_COMBINE=1 is introduced.
../src/xrt/drivers/ultraleap_v2/ulv2_driver.cpp: In function 'void* leap_input_loop(void*)':
../src/xrt/drivers/ultraleap_v2/ulv2_driver.cpp:290:29: warning: 'hi' may be used uninitialized in this function [-Wmaybe-uninitialized]
290 | int hi; // hand index
| ^~
It's not worth introducing a new error code for this case, just
return XRT_ERROR_VULKAN.
Fixes error with asserts disabled
../src/xrt/compositor/main/comp_compositor.c: In function 'compositor_mark_frame':
../src/xrt/compositor/main/comp_compositor.c:186:1: warning: control reaches end of non-void function [-Wreturn-type]
186 | }
| ^
In release builds asserts will be disabled, so we should try to handle the
assert cases reasonably.
This commit fixes several unused variable warnings when asserts are disabled.
The hardcoded value 32 was actually wrong and caused a warning
../src/xrt/state_trackers/oxr/oxr_input.c:668:9: warning: 'oxr_binding_find_bindings_from_key' accessing 256 bytes in a region of size 128 [-Wstringop-overflow=]
668 | oxr_binding_find_bindings_from_key(log, profile, act->act_key, binding_points, &num);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
It produces a warning
../src/xrt/state_trackers/gui/gui_scene_debug.c:141:9: warning: ‘igInputFloat3’ accessing 12 bytes in a region of size 4 [-Wstringop-overflow=]
141 | igInputFloat3(name, (float *) &value.x, "%+f", ImGuiInputTextFlags_ReadOnly);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
