Users may want to fall back to the CPU ASTC texture decoder due to hangs
and crashes that may be caused by keeping the GPU under compute heavy
loads for extended periods of time. This is especially the case in games
such as Astral Chain which make extensive use of ASTC textures.

lm: Demote LM guest logs to LOG_DEBUG

Guest logs are not very useful, as they are intended for use by the game developers during development. As such, they provide little meaning to be logged by yuzu and tend to overwhelm the log output at times.

configure_cpu_debug: Clarify settings behavior

common: fs: Use the normal directory iterator in *Recursively functions

common: logging: Restructure logging backend

Allows us to forward declare Common::FS::IOFile.

hid: Stub IsFirmwareUpdateAvailableForSixAxisSensor

This makes it clear that the disabled settings only take effect when CPU Accuracy is set to Debug Mode.

MSVC's implementation of recursive_directory_iterator throws an exception on an error despite a std::error_code being passed into its constructor. This is most likely a bug in MSVC's implementation since directory_iterator does not throw an exception on an error.

We can replace the usage of recursive_directory_iterator for now until MSVC fixes their implementation of it.

externals: Don't set FOUND or VERSION on LIBUSB

Fixes an issue where libusb.h wouldn't be found when building yuzu on
MSVC.

This only affects the "traditional" CMake pathway for linking libusb to
yuzu AKA MSVC. For autotools we still want to set these variables before
configuring SDL.

yuzu: main: Ensure enough space is available for RomFS dumping

Implement/Port Fastmem from Citra to Yuzu

This warns the user if there isn't enough free space to dump the entire RomFS to disk. It requires at least the size of the extracted RomFS + 1 GiB as a buffer of free space.

kernel: KLightConditionVariable: Update implementation to 12.x

This falls back to the old approach of using a virtual buffer.

Windows is untested, but this build should fix support for Windows < 10 v1803. However without fastmem support at all.

Workaround old headers and libraries shipped on MinGW.

In theory, if we have 2 MB continously mapped, this should save one layer of TLB.
Let's make it at least more likely by aligning the memory.

externals: Update SDL to 2f248a2a

cmake: Use autotools for libusb linking generally on GNU, and cleanup

Fix GCC undefined behavior sanitizer.

Updates the implementation of KLightConditionVariable to FW 12.x

hle: service: sm: Remove redundant session reservation, etc.

* Wrong alignment in u64 LOG_DEBUG -> memcpy.
* Huge shift exponent in stride calculation for linear buffer, unused result -> skipped.
* Large shift in buffer cache if word = 0, skip checking for set bits.

Non of those were critical, so this should not change any behavior.
At least with the assumption, that the last one used masking behavior, which always yield continuous_bits = 0.

- We were double-reserving, causing us to run out of sessions in Pokemon Sword & Shield.

- Pokemon Sword/Shield are still hitting this for some reason, causing an svcBreak.