GPU: Process command mode 5 (IncreaseOnce) differently from other commands

Accumulate all arguments before calling the desired method.

Note: Maybe we should do the same for the NonIncreasing mode?

GPU: Added some shader-related registers.

nvflinger: Remove superfluous buffer format check.

Shader address calculation depends on this value to some extent, we do not currently know what it being 0 entails.

Various heap fixes for libtransistor

Service/NIFM: convert to module

core: Move process creation out of global state.

Make Dynarmic the default CPU core

GPU: Intercept writes to the VERTEX_END_GL register.

This is the register that gets written after a game calls DrawArrays().

We should collect all GPU state and draw using our graphics API here.

FS: Make EnsureSaveData create the save data if it doesn't already exist.

CoreTiming: Unschedule the pending events when an Interface is destroyed

Vi: Signal the BufferQueue's Native Handle right after ReleaseBuffer is called

Service/Set: add more services

This prevents a thread starvation issue in Puyo Puyo Tetris.
We should hwtest this behavior and figure out where exactly this event is signaled.

Implemented the SaveData archive and MountSaveData.

There may be many CodeSets per Process, so it's wasteful and overcomplicated to store the program id in each of them.