Replace the previous OpenGL loader with a glad-generated 3.3 one

Add cam:u service function names to its function table

x64: Proper stack alignment in shader JIT function calls

video_core: Fix format specifiers warnings

Import Dolphin stack handling and register saving routines
Also removes the x86 parts from abi files

SVC: Advance time when calling GetSystemTick to escape busy-wait loops

Common: Fix MicroProfile compilation in MSVC2015

Pica: Add the vertex_offset register to the Pica registers map.

Only check src/ directory for whitespace errors in the pre-commit hook

Shader JIT: Fix SGE/SGEI NaN behavior

This allows importing of external libraries into externals/ without
having to reformat them. Unfortunately it also allows whitespace to be
introduced in files like the root CMakeLists.txt, but that is a small
downside compared to the tradeoff.

SGE was incorrectly emulated w.r.t. NaN behavior as the CMPSS SSE
instruction was used with NLT

GPU: Implemented register 0x22A PICA_REG_DRAW_VERTEX_OFFSET

This is the equivalent of the "first" parameter in glDrawArrays, it tells the GPU the vertex index at which to start rendering.

Register 0x22A doesn't affect indexed rendering.

Services/FS: Correctly tell the guest app whether a file was correctly opened or not

This gives us several niceties such as Sampler Objects, shader attribute
locations and Timer Queries.

The main advantage of switching to glad from glLoadGen is that, apart
from being actively maintained, it supports a customizable entrypoint
loader function, which makes it possible to also support OpenGL ES.

Cubic Ninja waited for the frame to end by spinning on a loop calling
GetSystemTick while doing nothing else. Since GetSystemTick doesn't
cause a reschedule (which advances time), this meant that very little
emulated time would pass inside that loop, causing the game to spend
most of the frame burning away CPU.

Rasterizer: Corrected the stencil implementation.

Closes #1067

Kernel: Fix wrong linear heap base on titles using newer kernels

Typo which sneaked in through review on #1025

Fix heap-management regressions

Shader FP compliance fixes

gl_rasterizer_cache: Cache textures based on hash in addition to address.

Shader JIT: Fix float to integer rounding in MOVA

MOVA converts new address register values from floats to integers using truncation

Should fix invalid read loops in some games

dyncom: Minor changes to CondPassed

Shader JIT: ifdef out reference to ifdef'd out shader_map

shader_map was only defined on x86 architectures, but was cleared on shutdown
with no ifdef protection. Ifdef this out so non-x86 architectures can be built.