[llvm] [BPF] Handle unreachable with a kfunc call (PR #131731)

[via llvm-commits]

================
@@ -726,6 +733,38 @@ SDValue BPFTargetLowering::LowerATOMIC_LOAD_STORE(SDValue Op,
   return Op;
 }
 
+SDValue BPFTargetLowering::LowerTRAP(SDValue Op, SelectionDAG &DAG) const {
+  MachineFunction &MF = DAG.getMachineFunction();
+  Function &F = MF.getFunction();
+  if (F.hasFnAttribute(Attribute::Naked))
+    return Op;
+
+  TargetLowering::CallLoweringInfo CLI(DAG);
+  SmallVector<SDValue> InVals;
+  SDNode *N = Op.getNode();
+  SDLoc DL(N);
+
+  Module *M = MF.getFunction().getParent();
+  FunctionType *FT = FunctionType::get(Type::getVoidTy(M->getContext()), false);
+  Function *UnreachableHelper = M->getFunction("__unreachable_helper");
+  if (!UnreachableHelper) {
+    Function *NewF = Function::Create(FT, GlobalValue::ExternalWeakLinkage,
+                                      "__unreachable_helper", M);
+    NewF->setDSOLocal(true);
+    NewF->setCallingConv(CallingConv::C);
----------------
eddyz87 wrote:

> basic block bb.1.flabel does not have terminator after the __unreachable_helper call

Basing on the comments for `TargetInstrInfo::analyzeBranch` absence of terminators means falltrhough.
But if that is so, how come call to `__unreachable_helper` does not clobber r1-r5 in this example?
>From llvm point of view there is a path `bb.0.entry` -> `bb.1.flabel` -> `bb.2.tlabel`, but register allocator somehow concludes that r1-r5 are intact after `bb.1.flabel`.

https://github.com/llvm/llvm-project/pull/131731