[llvm] [LLVM][AArch64] Improve big endian code generation for SVE BITCASTs. (PR #104769)

[Paul Walker via llvm-commits]

================
@@ -28877,7 +28885,20 @@ SDValue AArch64TargetLowering::getSVESafeBitCast(EVT VT, SDValue Op,
   if (InVT != PackedInVT)
     Op = DAG.getNode(AArch64ISD::REINTERPRET_CAST, DL, PackedInVT, Op);
 
-  Op = DAG.getNode(ISD::BITCAST, DL, PackedVT, Op);
+  if (Subtarget->isLittleEndian() ||
+      PackedVT.getScalarSizeInBits() == PackedInVT.getScalarSizeInBits())
+    Op = DAG.getNode(ISD::BITCAST, DL, PackedVT, Op);
+  else {
+    EVT PackedVTAsInt = PackedVT.changeTypeToInteger();
+    EVT PackedInVTAsInt = PackedInVT.changeTypeToInteger();
+
+    // Simulate the effect of casting through memory.
+    Op = DAG.getNode(ISD::BITCAST, DL, PackedInVTAsInt, Op);
+    Op = DAG.getNode(ISD::BSWAP, DL, PackedInVTAsInt, Op);
----------------
paulwalker-arm wrote:

I wasn't planning to.  The best big endian code generation is not the rational for the work.  I'm fixing bugs with how we're using uzp, uupk{lo,hi} and zip instructions and the fixes keep tripping over `getSVESafeBitCast` and the current big endian support so it's just easier for me to move the support along a little to clear the path for what I really want to fix.

https://github.com/llvm/llvm-project/pull/104769