[ARM] MVE: support QQPRRegClass and QQQQPRRegClass

[miyuki]

Summary

QQPRRegClass and QQQQPRRegClass are used by the
interleaving/deinterleaving loads/stores to represent sequences of
consecutive SIMD registers.

[simon_tatham]

Otherwise LGTM.

llvm/lib/Target/ARM/ARMISelLowering.cpp

1568 ↗

(On Diff #207293)

Comment nitpick: 4 to 8 consecutive NEON D registers, or 2 to 4 consecutive MVE Q registers.

[miyuki]

Fix a comment

llvm/trunk/lib/Target/ARM/ARMISelLowering.cpp

 
 /// getRegClassFor - Return the register class that should be used for the
 /// specified value type.
 const TargetRegisterClass *
 ARMTargetLowering::getRegClassFor(MVT VT, bool isDivergent) const {
   (void)isDivergent;
   // Map v4i64 to QQ registers but do not make the type legal. Similarly map
   // v8i64 to QQQQ registers. v4i64 and v8i64 are only used for REG_SEQUENCE to
-  // load / store 4 to 8 consecutive D registers.
-  if (Subtarget->hasNEON()) {
+  // load / store 4 to 8 consecutive NEON D registers, or 2 to 4 consecutive
+  // MVE Q registers.
+  if (Subtarget->hasNEON() || Subtarget->hasMVEIntegerOps()) {
     if (VT == MVT::v4i64)
       return &ARM::QQPRRegClass;
     if (VT == MVT::v8i64)
       return &ARM::QQQQPRRegClass;
   }
   return TargetLowering::getRegClassFor(VT);
 }