Improvements to rasterizer cache

file_util: Remove redundant duplicate return in GetPathWithoutTop()

video_core: Use nested namespaces where applicable

common: Remove synchronized_wrapper.h

file_util: Use an enum class for GetUserPath()

GPU: Implement the NVGPU_IOCTL_CHANNEL_KICKOFF_PB ioctl2 command.

This behaves quite similarly to the SubmitGPFIFO command. Referenced from Ryujinx.
Many thanks to @gdkchan for investigating this!

Instead of using an unsigned int as a parameter and expecting a user to
always pass in the correct values, we can just convert the enum into an
enum class and use that type as the parameter type instead, which makes
the interface more type safe.

We also get rid of the bookkeeping "NUM_" element in the enum by just
using an unordered map. This function is generally low-frequency in
terms of calls (and I'd hope so, considering otherwise would mean we're
slamming the disk with IO all the time) so I'd consider this acceptable
in this case.

Given both operands are the same type, there won't be an issue with
overload selection that requires making this explicit.

This is entirely unused in the codebase.

partition_filesystem, vfs_real: Minor changes

arm_interface: Remove unused tls_address member of ThreadContext

externals: Update dynarmic to 7ea1241

tests/arm_test_common: Minor changes

gl_shader_manager: Remove unimplemented function prototype

file_sys/errors: Remove redundant object constructor calls

gpu: Rename Get3DEngine() to Maxwell3D()

CPU: Save and restore the TPIDR_EL0 system register on every context switch

vfs: Minor changes

Resolves an issue with TPIDR setting being erroneously removed in the
dead code pass.

Loader: Only print the module names and addresses if they actually exist.

Given we're already constructing the error code, we don't need to call
the constructor inside of it.

We already return by value, so we don't explicitly need to make the
copy.

Avoids unnecessary atomic increment and decrement operations.

This is a little bit more self-documenting on what is being done here.

We can simply use std::clamp() here, instead of using an equivalent
with std::max() and std::min().

Given the data is intended to be directly written, there's no need to
take the std::vector by value and copy the data.

Provides the same behavior, but with less writing

vfs: Amend constness on pointers in WriteBytes() and WriteArrays() member functions to be const qualified

These functions don't modify the data being pointed to, so these can be
pointers to const data

Note that there's currently a dynarmic bug preventing this register from being written.

logging: Use std::string_view where applicable