[llvm] [AArch64][GlobalISel] Allow selecting FPR index loads. (PR #143835)

[David Green via llvm-commits]

davemgreen wrote:

It is not doing an fpext if that is what you mean (a half->float convert). Just an anyext (s16->s32, it doesn't matter what is in the upper bits). Which sounds like what we already have. AArch64 doesn't have anything that would need a G_FPEXT_LOAD if it performed an fp convert.

Some of this comes from the way we legalize selects at the moment, we don't have a way to tell apart fp and integer variants. That can be fixed in other ways, but I figured it was worth fixing this for G_INDEXED_LOADS anyway, in case they come up for other reasons.

https://github.com/llvm/llvm-project/pull/143835