[LoongArch64] A plan for amending the LoongArch64‘s port

[shushanhf]

Description

After the supporting for LoongArch64 finished by #59561 ,
there are still some TODOs for LA that should be implemented.

Here creating a new issue to track these.

(1) Fixing all the bugs anytime;
(2) Amend the LoongArch64's ABI;
(3) Enable the promote feature;
(4) Pushing the crossgen2 for LA;
(5) Supporting the SIMD/intrinsic for LA;
(6) ...

The CoreRoot for LoongArch64 will be updated here

Reproduction Steps

Testing all the test-cases of the runtime on LoongArch64's Linux.

Expected behavior

The test cases should pass.

Actual behavior

Failed.

Regression?

No response

Known Workarounds

Using the SDK6-LoongArch64 which published by Loongson based on the offical-SDK6.

Configuration

LoongArch64’s Linux

Other information

No response

I couldn't figure out the best area label to add to this issue. If you have write-permissions please help me learn by adding exactly one area label.

[shushanhf]

@jkotas @BruceForstall

we will push some PRs to amend the LA's port after we had fixed some test cases on LA-linux.

[huoyaoyuan]

Supporting the SIMD/intrinsic for LA

I've read the instruction spec for LA, and found some instructions interesting. As a brand new architecture, we may be concerned about exposing intrinsics in public API. However, we can implement an internal intrinsic class first and use in BitOperations etc.

[shushanhf]

Supporting the SIMD/intrinsic for LA

I've read the instruction spec for LA, and found some instructions interesting. As a bread new architecture, we may be concerned about exposing intrinsics in public API. However, we can implement an internal intrinsic class first and use in BitOperations etc.

Thanks.