[llvm] AArch64: align pair-wise spills on WoS to 16-byte (PR #166902)

[Martin Storsjö via llvm-commits]

================
@@ -1568,8 +1569,14 @@ static bool invalidateWindowsRegisterPairing(unsigned Reg1, unsigned Reg2,
     return true;
   if (!NeedsWinCFI)
     return false;
+  // ARM64EC introduced `save_any_regp` which expects 16-byte alignment.
+  // Accomodate that by ensuring that we re-align to 16-bytes when doing paired
----------------
mstorsjo wrote:

Then again, this does regress the generated code for generic arm64 targets that don't use this calling convention (some prologues end up with one store more than before), so I wonder what @efriedma-quic thinks about that.

https://github.com/llvm/llvm-project/pull/166902