[llvm] [AArch64][SME] Allow memory operations lowering to custom SME functions. (PR #79263)

[Sander de Smalen via llvm-commits]

================
@@ -76,15 +82,85 @@ SDValue AArch64SelectionDAGInfo::EmitMOPS(AArch64ISD::NodeType SDOpcode,
   }
 }
 
+SDValue AArch64SelectionDAGInfo::EmitSpecializedLibcall(
----------------
sdesmalen-arm wrote:

I see how you changed the handling of different return types, but we still have the code for the 'size' that is uniform for the three case-statements at the end, but which isn't generic for other libcalls (if we ever add them). It's probably better to just rename this function to what it's meant to be: `EmitStreamingCompatibleMemLibCall` and then fix `RetTy` to be `PointerType::getUnqual()`, rather than setting it in each case statement.

https://github.com/llvm/llvm-project/pull/79263