Drop the std::list hack to allocate memory indefinitely.

Instead use a custom allocator that keeps references valid until
destruction. This allocates fixed chunks of memory and puts pointers in
a free list. When an allocation is no longer used put it back to the
free list, this doesn't heap allocate because std::vector doesn't change
the capacity. If the free list is empty, allocate a new chunk.

Most overlaps in the buffer cache only contain one mapped address.
We can avoid close to all heap allocations once the buffer cache is
warmed up by using a small_vector with a stack size of one.

Instead of using boost::icl::interval_map for caching, use
boost::intrusive::set. interval_map is intended as a container where the
keys can overlap with one another; we don't need this for caching
buffers and a std::set-like data structure that allows us to search with
lower_bound is enough.

Removing shared pointers is a first step to be able to use intrusive
objects and keep allocations close to one another in memory.

Minor style changes. Mostly done so I avoid editing it while doing other
changes.

hid: Clear keyboard states & fix logic issue

FS: Improve emulation of device saves

nv_flinger: Use enum for pixel format instead of u32

DmaPusher: Remove dead code in step

vk_rasterizer: Match OpenGL's FlushAndInvalidate behavior

Match OpenGL's behavior. This can fix or simplify bisecting issues on
Vulkan.

Frontend: Remove tracking for context wrapper

audio_renderer: Better voice mixing and 6 channel downmixing

Improve time zone support

shader_ir: Add separate instructions for ordered and unordered comparisons and fix NE on GLSL

Previously we never cleared the states of the entries and the key would stay held down, also looping over the key bytes for each key lead to setting every bit for the key state instead of the key we wanted

vk_graphics_pipeline: Implement rasterizer_enable on Vulkan

Stub hid:SendKeyboardLockKeyEvent

Supersedes #3738 and #3321

Needed for Puchikon 4 SmileBASIC 1.0.0

CI: Don't check clang format on early access builds

"Not equal" operators on GLSL seem to behave as unordered when we expect
an ordered comparison.

Manually emulate this checking for LGE values (numbers, not-NaNs).

externals: Cmake version checks