Update Vulkan-Headers to 1.4.316 #951
Conversation
MarijnS95
changed the title
|
With the I don't think we can address this in the current release, but I do think we can solve #343 by adding a new (This is not "entirely" valid in Rust terms as I don't think feature flags "allow" you to opt out of semver compatibility even if you name them as such, and I'm also not sure how often we have to rely on it, but the current "just generate/wrap it anyway" approach does lock us out of making non-breaking releases with |
MarijnS95
changed the title
MarijnS95
changed the title
MarijnS95
changed the title
| pub cmd_write_buffer_marker2_amd: PFN_vkCmdWriteBufferMarker2AMD, | ||
| pub get_queue_checkpoint_data2_nv: PFN_vkGetQueueCheckpointData2NV, |
There was a problem hiding this comment.
Technically a semver-breaking change too. These functions were provided by VK_AMD_buffer_marker or VK_NV_device_diagnostic_checkpoints respectively, but vk.xml previously exposed them under the VK_KHR_synchronization2 extension, and so did our generator.
MarijnS95
changed the title
ash/src/device.rs
Outdated
| /// <https://registry.khronos.org/vulkan/specs/1.3-extensions/man/html/vkMapMemory2.html> | ||
| #[inline] | ||
| #[doc(alias = "vkMapMemory2")] | ||
| pub unsafe fn map_memory2( | ||
| &self, | ||
| memory_map_info: &vk::MemoryMapInfo<'_>, | ||
| ) -> VkResult<*mut ffi::c_void> { | ||
| let mut data = mem::MaybeUninit::uninit(); | ||
| (self.device_fn_1_4.map_memory2)(self.handle, memory_map_info, data.as_mut_ptr()) | ||
| .assume_init_on_success(data) | ||
| } | ||
|
|
||
| /// <https://registry.khronos.org/vulkan/specs/1.3-extensions/man/html/vkUnmapMemory2.html> | ||
| #[inline] | ||
| #[doc(alias = "vkUnmapMemory2")] | ||
| pub unsafe fn unmap_memory2( | ||
| &self, | ||
| memory_unmap_info: &vk::MemoryUnmapInfo<'_>, | ||
| ) -> VkResult<()> { | ||
| (self.device_fn_1_4.unmap_memory2)(self.handle, memory_unmap_info).result() | ||
| } |
There was a problem hiding this comment.
Of all the stabilized extensions, only VK_KHR_map_memory2 was missing a wrapper implementation in Ash. I'm considering to retroactively add it.
| // XXX: The below two belong in a different extension! | ||
|
|
||
| /// <https://registry.khronos.org/vulkan/specs/1.3-extensions/man/html/vkCmdSetDescriptorBufferOffsets2EXT.html> | ||
| #[inline] | ||
| #[doc(alias = "vkCmdSetDescriptorBufferOffsets2EXT")] | ||
| pub unsafe fn cmd_set_descriptor_buffer_offsets2( |
There was a problem hiding this comment.
In vkCmdSetDescriptorBufferOffsets2EXT and vkCmdBindDescriptorBufferEmbeddedSamplers2EXT the <require> block sits inside VK_KHR_maintenance6 with an extra depends="VK_EXT_descriptor_buffer". It's a bit ugly to have the functions like this I think, given that they're EXT and not KHR. There are more such cases though.
But the Vulkan documentation also says it's clearly provided by VK_KHR_maintenance6 if VK_EXT_descriptor_buffer is also available: https://registry.khronos.org/vulkan/specs/latest/man/html/VK_KHR_maintenance6.html#_new_commands
And nothing is listed for VK_EXT_descriptor_buffer: https://registry.khronos.org/vulkan/specs/latest/man/html/VK_EXT_descriptor_buffer.html#_new_commands
|
Marking as draft to set expectations: this will be rebased/updated and merged after we've gone through all generator churn, which is important to get through first. |
This comment was marked as outdated.
This comment was marked as outdated.
MarijnS95
changed the title
MarijnS95
force-pushed
the
update
branch
2 times, most recently
from
9d780df to
cc89f2a
Compare
2 participants
No description provided.