[llvm] [RISCV][VLOPT] Add vfsqrt/vfrsqrt7 instruction to isSupportInstr (PR #127462)

[Luke Lau via llvm-commits]

================
@@ -5069,3 +5069,63 @@ define <vscale x 4 x float> @vfwmaccbf16_vf(<vscale x 4 x float> %a, bfloat %b,
   %2 = call <vscale x 4 x float> @llvm.riscv.vfadd(<vscale x 4 x float> poison, <vscale x 4 x float> %1, <vscale x 4 x float> %d, iXLen 7, iXLen %vl)
   ret <vscale x 4 x float> %2
 }
+
+define <vscale x 4 x half> @vfsqrt(<vscale x 4 x half> %a) {
----------------
lukel97 wrote:

Do we need +zvfh or +zvfbfmin? Can we just test it with floats + fadd e.g. how `@vfwmaccbf16_vf` does it

https://github.com/llvm/llvm-project/pull/127462