Implement fast path for AES crypto instructions on Arm64

[gdkchan]

As the title says. This greatly improves the performance of AES operations when using JIT on Arm64 platforms. It is most notable on Animal Crossing when the game is saving, since it apparently encrypts the save data, and before this change that would cause noticeable stutters.

Closes #5263.

[github-actions]

Download the artifacts for this pull request:

• ryujinx-Release-1.1.0+0177526-linux_x64
• ryujinx-Release-1.1.0+0177526-win_x64

Experimental GUI (Avalonia)

• ava-ryujinx-Release-1.1.0+0177526-linux_x64
• ava-ryujinx-Release-1.1.0+0177526-macos_universal
• ava-ryujinx-Release-1.1.0+0177526-win_x64

GUI-less (SDL2)

• sdl2-ryujinx-headless-Release-1.1.0+0177526-linux_x64
• sdl2-ryujinx-headless-Release-1.1.0+0177526-win_x64

Only for Developers

• ava-ryujinx-Debug-1.1.0+0177526-linux_x64
• ava-ryujinx-Debug-1.1.0+0177526-macos_universal
• ava-ryujinx-Debug-1.1.0+0177526-win_x64
• ryujinx-Debug-1.1.0+0177526-linux_x64
• ryujinx-Debug-1.1.0+0177526-win_x64
• sdl2-ryujinx-headless-Debug-1.1.0+0177526-linux_x64
• sdl2-ryujinx-headless-Debug-1.1.0+0177526-win_x64

[riperiperi]

Nice. 😌

[gdkchan]

I tested Animal Crossing on macOS and it's working properly (no more heavy stutter when saving with JIT). I also ran the Crypto test (and Crypto32 too) on a Raspberry Pi 4, they all pass, but unfortunately the Raspberry Pi CPU does not support the AES extension, so it was still taking the slow paths there. It would be nice if someone could run the tests on Arm64 Linux with a CPU that supports AES. Almost none of the CPU tests run on Arm64 macOS right now AFAIK due to Unicorn issues.

[aletrevi99]

this pr also cuts loading times by more than half in acnh! used to take around 10 seconds to load a new room for example, now around 3/4 seconds

[jduncanator]

Tests are passing correctly on an ARM64 Linux target here:

Features        : fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm lrcpc dcpop asimddp ssbs

Test run for src/Ryujinx.Tests/bin/Debug/net7.0/Ryujinx.Tests.dll (.NETCoreApp,Version=v7.0)
Microsoft (R) Test Execution Command Line Tool Version 17.6.0 (arm64)
Copyright (c) Microsoft Corporation.  All rights reserved.

Starting test execution, please wait...
A total of 1 test files matched the specified pattern.
NUnit Adapter 4.1.0.0: Test execution started
Running selected tests in src/Ryujinx.Tests/bin/Debug/net7.0/Ryujinx.Tests.dll
   NUnit3TestExecutor discovered 24 of 24 NUnit test cases using Current Discovery mode, Non-Explicit run
  Passed Aesd_V(uint.MinValue,1,8888791084251240037UL,7166475444035995485UL,4173465908070920181UL,9943920176957912833UL,10217182921803957948UL,10328748588515655053UL) [395 ms]
  Passed Aesd_V(uint.MinValue,1,8888791084251240037UL,7166475444035995485UL,4173465908070920181UL,15351740853312108513UL,10217182921803957948UL,10328748588515655053UL) [4 ms]
  Passed Aesd_V(uint.MinValue,1,8888791084251240037UL,7166475444035995485UL,8827430708575053269UL,9943920176957912833UL,10217182921803957948UL,10328748588515655053UL) [2 ms]
  Passed Aesd_V(uint.MinValue,1,8888791084251240037UL,7166475444035995485UL,8827430708575053269UL,15351740853312108513UL,10217182921803957948UL,10328748588515655053UL) [2 ms]
  Passed Aese_V(uint.MinValue,1,8888791084251240037UL,7166475444035995485UL,14273691190541683539UL,4127970860932264262UL,10345507550822998093UL,5492616109554313292UL) [3 ms]
  Passed Aese_V(uint.MinValue,1,8888791084251240037UL,7166475444035995485UL,14273691190541683539UL,1690253263847695179UL,10345507550822998093UL,5492616109554313292UL) [2 ms]
  Passed Aese_V(uint.MinValue,1,8888791084251240037UL,7166475444035995485UL,10484708430689203144UL,4127970860932264262UL,10345507550822998093UL,5492616109554313292UL) [2 ms]
  Passed Aese_V(uint.MinValue,1,8888791084251240037UL,7166475444035995485UL,10484708430689203144UL,1690253263847695179UL,10345507550822998093UL,5492616109554313292UL) [2 ms]
  Passed Aesimc_V(uint.MinValue,1,10217183059242911420UL,10328748588515655053UL,15435426261652692654UL,17206508408432187255UL) [3 ms]
  Passed Aesimc_V(uint.MinValue,uint.MinValue,10217183059242911420UL,10328748588515655053UL,15435426261652692654UL,17206508408432187255UL) [2 ms]
  Passed Aesmc_V(uint.MinValue,1,7096106647905438152UL,3105288063291732965UL,8888791084251240037UL,7166475444035995485UL) [3 ms]
  Passed Aesmc_V(uint.MinValue,uint.MinValue,7096106647905438152UL,3105288063291732965UL,8888791084251240037UL,7166475444035995485UL) [2 ms]
  Passed Aesd_V(uint.MinValue,2,8888791084251240037UL,7166475444035995485UL,2290942773153415507UL,11919168958138124598UL,10217182921803957948UL,10328748588515655053UL) [16 ms]
  Passed Aesd_V(uint.MinValue,2,8888791084251240037UL,7166475444035995485UL,2290942773153415507UL,9519032407040401072UL,10217182921803957948UL,10328748588515655053UL) [8 ms]
  Passed Aesd_V(uint.MinValue,2,8888791084251240037UL,7166475444035995485UL,4689522466461389379UL,11919168958138124598UL,10217182921803957948UL,10328748588515655053UL) [4 ms]
  Passed Aesd_V(uint.MinValue,2,8888791084251240037UL,7166475444035995485UL,4689522466461389379UL,9519032407040401072UL,10217182921803957948UL,10328748588515655053UL) [4 ms]
  Passed Aese_V(uint.MinValue,2,8888791084251240037UL,7166475444035995485UL,16289406851345172903UL,13102951385261698891UL,10345507550822998093UL,5492616109554313292UL) [5 ms]
  Passed Aese_V(uint.MinValue,2,8888791084251240037UL,7166475444035995485UL,16289406851345172903UL,14608110044635135832UL,10345507550822998093UL,5492616109554313292UL) [4 ms]
NUnit Adapter 4.1.0.0: Test execution complete
  Passed Aese_V(uint.MinValue,2,8888791084251240037UL,7166475444035995485UL,16515115464934646839UL,13102951385261698891UL,10345507550822998093UL,5492616109554313292UL) [4 ms]
  Passed Aese_V(uint.MinValue,2,8888791084251240037UL,7166475444035995485UL,16515115464934646839UL,14608110044635135832UL,10345507550822998093UL,5492616109554313292UL) [4 ms]
  Passed Aesimc_V(uint.MinValue,2,10217183059242911420UL,10328748588515655053UL,15435426261652692654UL,17206508408432187255UL) [5 ms]
  Passed Aesimc_V(uint.MinValue,uint.MinValue,10217183059242911420UL,10328748588515655053UL,15435426261652692654UL,17206508408432187255UL) [5 ms]
  Passed Aesmc_V(uint.MinValue,2,7096106647905438152UL,3105288063291732965UL,8888791084251240037UL,7166475444035995485UL) [6 ms]
  Passed Aesmc_V(uint.MinValue,uint.MinValue,7096106647905438152UL,3105288063291732965UL,8888791084251240037UL,7166475444035995485UL) [4 ms]

Test Run Successful.
Total tests: 24
     Passed: 24
 Total time: 20.9817 Seconds

[fnsmake]

run smooth！ thank u for your work~