CI: auto-publish Android releases

Use `toUtf8()` for string passed to DBus

core_timing: Remove GetCurrentTimerResolution in CoreTiming loop

Add support for VK_KHR_image_format_list

kernel: Synchronize

Other programs may change this value, but if thousands of syscalls in this loop is undesirable, then we can just set this once.

android: Rework MemoryUtil

Memory Tracking: Add mechanism to register small writes when gpu page is contested by GPU

fsmitm_romfsbuild: avoid full path lookups

Revert "texture_cache: Fix incorrect logic for AccelerateDMA"

This reverts commit 1fc47361.

Minor cleanup in BufferCacheRuntime::ReserveNullBuffer

vulkan: Improvements to macOS surface creation

general: Misc changes that did not deserve their own PRs

sink_stream: Resolve heap buffer corruption due to out of bounds write

input_common: Tune mouse controls

It is probably not correct to create a surface on a non-VulkanSurface window.
On macOS this causes a preferences crash due to missing CAMetalLayer.

As far as I can tell, there is no reason to OR this bit in separately.

* Many times the format itself wouldn't have been added to the list causing device losses for nvidia GPUs

* Also account for ASTC acceleration storage views

Co-authored-by: The yuzu Community <noreply-fake@community.yuzu-emu.org>

Texture cache: Fix YFC regression due to code testing

* Switch OpenJDK runtime to Eclipse Temurin (AdoptOpenJDK has rebranded
to Eclipse Temurin)
* Fetch submodules using full clones instead of shallow clones

Allows for implicit conversion to std::span<T>.

AccelerateDMA: Don't accelerate 3D texture DMA operations

Uses string templates and rounds up memory amount for potentially inaccurate checks now

Also, remove the use of ScratchBuffer when upmixing, as other channels may not be initialized with zeroed out data.

android: Suppress a known incompatibility

x64: Make use of monitorx instructions for power efficient sleeps (AMD)