[AArch64] Optimize more floating-point round+convert combinations into fcvt instructions#170018

valadaptive wants to merge 4 commits intollvm:mainllvm/llvm-project:mainfrom

valadaptive:aarch64-round-convertvaladaptive/llvm-project:aarch64-round-convertCopy head branch name to clipboard