[llvm] [LLVM][AArch64]Use load/store with consecutive registers in SME2 or S… (PR #77665)

[Sander de Smalen via llvm-commits]

================
@@ -3065,19 +3074,40 @@ bool AArch64FrameLowering::spillCalleeSavedRegisters(
       std::swap(Reg1, Reg2);
       std::swap(FrameIdxReg1, FrameIdxReg2);
     }
+
+    unsigned PairRegs;
+    unsigned PnReg;
+    if (RPI.isPaired() && RPI.isScalable()) {
+      PairRegs = AArch64::Z0_Z1 + (RPI.Reg1 - AArch64::Z0);
+      if (!PtrueCreated) {
+        PtrueCreated = true;
+        // Any one of predicate-as-count will be free to use
+        // This can be replaced in the future if needed
+        PnReg = AArch64::PN8;
----------------
sdesmalen-arm wrote:

I'm a little concerned with just blindly picking `pn8` here, because this may not match the given calling convention (if someone would choose to use a different one from the standard SVE calling convention), defined in AArch64CallingConvention.td.

Can you create a function that finds a suitable *caller*-saved register instead?

https://github.com/llvm/llvm-project/pull/77665