[llvm] [win][x64] Fix import call optimization for calls to dllimports and global function pointers (PR #160604)

[Daniel Paoliello via llvm-commits]

================
@@ -2551,13 +2552,19 @@ void X86AsmPrinter::emitInstruction(const MachineInstr *MI) {
       emitLabelAndRecordForImportCallOptimization(
           IMAGE_RETPOLINE_AMD64_IMPORT_CALL);
 
-      MCInst TmpInst;
-      MCInstLowering.Lower(MI, TmpInst);
-
       // For Import Call Optimization to work, we need a the call instruction
       // with a rex prefix, and a 5-byte nop after the call instruction.
       EmitAndCountInstruction(MCInstBuilder(X86::REX64_PREFIX));
-      emitCallInstruction(TmpInst);
+      // Following the pattern of how dllimport calls are lowered in FastISel.
----------------
dpaoliello wrote:

Fair enough. Added a new `CALL64_ImpCall` pseudo and lowered it to CALL64m in X86ExpandPseudo.

https://github.com/llvm/llvm-project/pull/160604