[llvm] [X86]Support lowering for APX Promoted SHA/MOVDIR/CRC32/INVPCID instructions (PR #76786)

[Shengchen Kan via llvm-commits]

================
@@ -3046,22 +3046,24 @@ bool X86FastISel::fastLowerIntrinsicCall(const IntrinsicInst *II) {
     switch (II->getIntrinsicID()) {
     default:
       llvm_unreachable("Unexpected intrinsic.");
+#define GET_EGPR_IF_ENABLED(OPC) Subtarget->hasEGPR() ? OPC##_EVEX : OPC;
     case Intrinsic::x86_sse42_crc32_32_8:
-      Opc = X86::CRC32r32r8;
+      Opc = GET_EGPR_IF_ENABLED(X86::CRC32r32r8);
       RC = &X86::GR32RegClass;
       break;
     case Intrinsic::x86_sse42_crc32_32_16:
-      Opc = X86::CRC32r32r16;
+      Opc = GET_EGPR_IF_ENABLED(X86::CRC32r32r16);
----------------
KanRobert wrote:

Maybe we don't need to select the EVEX variant for fast isel, which is used for O0 only. How do you think @phoebewang @topperc ?

https://github.com/llvm/llvm-project/pull/76786