[llvm] (WIP) [AArch64] Make use of byte FPR stores for bytes extracted from vectors (PR #134117)

[Benjamin Maxwell via llvm-commits]

================
@@ -107,94 +107,109 @@ define <16 x i8> @test_compress_v16i8(<16 x i8> %vec, <16 x i1> %mask) {
 ; CHECK-NEXT:    sub sp, sp, #16
----------------
MacDue wrote:

Updated to prefer st1.b in cases like this.

https://github.com/llvm/llvm-project/pull/134117