The kernel does the equivalent of the following check before proceeding:

if (address + 0x8000000000 < 0x7FFFE00000) {
    return ERR_INVALID_MEMORY_STATE;
}

which is essentially what our IsKernelVirtualAddress() function does. So
we should also be checking for this.

The kernel also checks if the given input addresses are 4-byte aligned,
however our Mutex::TryAcquire() and Mutex::Release() functions already
handle this, so we don't need to add code for this case.

Optimized Texture Swizzling

service/vi: Replace includes with forward declarations where applicable

kernel/errors: Amend error code for ERR_NOT_FOUND

audio_core/time_stretch: Silence truncation warnings in Process()

gl_shader_decompiler: Get rid of variable shadowing within LEA instructions

services/sm: Amend error code constants

cubeb_sink: Get rid of variable shadowing within CubebSink's constructor

kernel/svc: Handle a few error cases within memory-related functions

cubeb_sink: Correct context name in ListCubebSinkDevices()

Port #4192 from Citra: "svc: change unknown to thread in CreateThread"

Port #4171 from Citra: "Tests: Remove glad test OS X work-around"

Port #4182 from Citra: "Prefix all size_t with std::"

Implement RenderTargetFormat::BGR5A1_UNORM

game_list_p: Take map iterator contents by const reference 

yuzu/util: Antialias game list compatibility pixmaps

Implement ASTC_2D_8X8 

We pass a hint to the QPainter instance that we want anti-aliasing on
the compatibility icons, which prevents the circles from looking fairly
jagged, and actually makes them look circular.

Adds a missing 'i' character that was missing in compatibility.

We don't need to copy the whole struct in this instance, we can just
utilize a reference instead.

Shaders: Implemented multiple-word loads and stores to and from attribute memory.

GPU: Basic implementation of the Kepler Inline Memory engine (p2mf).

This seems to be an optimization performed by nouveau.

The parameter of the lambda was shadowing the variable that was being
assigned to.

This ain't Citra.

Courtesy of @ogniK5377.

This also moves them into the cpp file and limits the visibility to
where they're directly used. It also gets rid of unused or duplicate
error codes.

The kernel caps the size limit of shared memory to 8589930496 bytes (or
(1GB - 512 bytes) * 8), so approximately 8GB, where every GB has a 512
byte sector taken off of it.

It also ensures the shared memory is created with either read or
read/write permissions for both permission types passed in, allowing the
remote permissions to also be set as "don't care".

kernel/svc: Sanitize addresses, permissions, and sizes within svcMapSharedMemory() and svcUnmapSharedMemory()

Part of the checking done by the kernel is to check if the given
address and size are 4KB aligned, as well as checking if the size isn't
zero. It also only allows mapping shared memory as readable or
read/write, but nothing else, and so we shouldn't allow mapping as
anything else either.

These variables are already defined within an outer scope.

The kernel checks if the addresses and given size is 4KB aligned before
continuing onwards to map the memory.

The kernel checks if the given size is a multiple of 2MB and <= to 4GB
before going ahead and attempting to allocate that much memory.

kernel/thread: Include thread-related enums within the kernel namespace

service: Use nested namespace specifiers where applicable

The SoundTouch API only accepts uint amount of samples.