[llvm] AMDGPU/GlobalISel: Switch to extended LLTs (PR #196522)

[Petar Avramovic via llvm-commits]

petar-avramovic wrote:

Most of the changes in common code are using LLT::integer instead of LLT::scalar for "integer opcodes" which was commonly used in aarch64 switch in https://github.com/llvm/llvm-project/pull/177158

changes affecting arm64 are in https://github.com/llvm/llvm-project/pull/196399

I still need to go though Legalizer changes, some cases still use scalar LLT and changing to explicit type might avoid the need change common code but did not yet go through all of it.


https://github.com/llvm/llvm-project/pull/196522