VideoCore: Fix OGL cache invalidation.

input_common: Create an update engine

Avoid OOB array access reading passthrough attr mask

YFC 1.5 extended the size of the varying mask used to hold passthrough attrs without considering this

renderer_vulkan: disable clock boost on unvalidated devices

vulkan_device: avoid attempt to access empty optional

opengl: Sanitize antialiasing config

video_core/vulkan: Fixed loading of Vulkan driver pipeline cache

The header size of the Vulkan driver pipeline cache files was incorrectly in PipelineCache::LoadVulkanPipelineCache, for which the pipeline cache wasn't read correctly and got invalidated on each load.

Port over several shader-compiler fixes from skyline

externals: update dynarmic, xbyak

net: Silently translate ETIMEDOUT network error

vulkan: implement 'turbo mode' clock booster 

VideoCore: Implement Cached Writes, use fastmem for reading GPU memory and eliminate old stuffs

MacroHLE: eliminate 2 rushed macros.

renderer_vulkan: implement fallback path for null buffer descriptors

vulkan_common: unify VK_EXT_debug_utils and selection of validation layer

The existing implementation only supports 64 invoc-per-subgroup GPUs, and misbehaves on adreno when invocations need to be split into 4 emulated subgroups.

Reuses most of the existing code for generating the gl_Layer passthrough. Fixes geometry in Nier: Automata on GPUs without HW passthrough support.

We can take advantage of SSBO addresses being passed in a constant bufer to account for the extra alignment requirements in the shader itself.

Adreno drivers will crash compiling geometry shaders if the input position is not wrapped in a gl_in struct.

video_core/vulkan: Implemented `VkPipelineCache` to store Vulkan pipelines

So that old cache can get deleted when the cache version changes and does not grow infinitely