1include::meta/VK_EXT_display_control.txt[] 2 3*Last Modified Date*:: 4 2016-12-13 5*IP Status*:: 6 No known IP claims. 7*Contributors*:: 8 - Pierre Boudier, NVIDIA 9 - James Jones, NVIDIA 10 - Damien Leone, NVIDIA 11 - Pierre-Loup Griffais, Valve 12 - Daniel Vetter, Intel 13 14This extension defines a set of utility functions for use with the 15`<<VK_KHR_display>>` and `<<VK_KHR_display_swapchain>>` extensions. 16 17=== New Enum Constants 18 19 * Extending elink:VkStructureType: 20 ** ename:VK_STRUCTURE_TYPE_DISPLAY_POWER_INFO_EXT 21 ** ename:VK_STRUCTURE_TYPE_DEVICE_EVENT_INFO_EXT 22 ** ename:VK_STRUCTURE_TYPE_DISPLAY_EVENT_INFO_EXT 23 ** ename:VK_STRUCTURE_TYPE_SWAPCHAIN_COUNTER_CREATE_INFO_EXT 24 25=== New Enums 26 27 * elink:VkDisplayPowerStateEXT 28 * elink:VkDeviceEventTypeEXT 29 * elink:VkDisplayEventTypeEXT 30 31=== New Structures 32 33 * slink:VkDisplayPowerInfoEXT 34 * slink:VkDeviceEventInfoEXT 35 * slink:VkDisplayEventInfoEXT 36 * slink:VkSwapchainCounterCreateInfoEXT 37 38=== New Functions 39 40 * flink:vkDisplayPowerControlEXT 41 * flink:vkRegisterDeviceEventEXT 42 * flink:vkRegisterDisplayEventEXT 43 * flink:vkGetSwapchainCounterEXT 44 45=== Issues 46 471) Should this extension add an explicit "`WaitForVsync`" API or a fence 48signaled at vsync that the application can wait on? 49 50*RESOLVED*: A fence. 51A separate API could later be provided that allows exporting the fence to a 52native object that could be inserted into standard run loops on POSIX and 53Windows systems. 54 552) Should callbacks be added for a vsync event, or in general to monitor 56events in Vulkan? 57 58*RESOLVED*: No, fences should be used. 59Some events are generated by interrupts which are managed in the kernel. 60In order to use a callback provided by the application, drivers would need 61to have the userspace driver spawn threads that would wait on the kernel 62event, and hence the callbacks could be difficult for the application to 63synchronize with its other work given they would arrive on a foreign thread. 64 653) Should vblank or scanline events be exposed? 66 67*RESOLVED*: Vblank events. 68Scanline events could be added by a separate extension, but the latency of 69processing an interrupt and waking up a userspace event is high enough that 70the accuracy of a scanline event would be rather low. 71Further, per-scanline interrupts are not supported by all hardware. 72 73=== Version History 74 75 * Revision 1, 2016-12-13 (James Jones) 76 - Initial draft 77