[llvm] [LoongArch] Enable callee saved register optimization (PR #226809)

[Shreeyash Pandey via llvm-commits]

bojle wrote:

Hi @lrzlin. I have a similar patch for aarch64 (see https://github.com/llvm/llvm-project/pull/214773/). Interestingly, I observed perf improvements on perlbench (SPEC2017). Additionally, I think this could also lead to compile time regressions like 214773 did (see https://github.com/llvm/llvm-project/pull/214773#issuecomment-5238390969).

https://github.com/llvm/llvm-project/pull/226809