[llvm] [AMDGPU] Update base addr of dyn alloca considering GrowingUp stack (PR #119822)

[Sergei Barannikov via llvm-commits]

================
@@ -4037,31 +4038,35 @@ SDValue SITargetLowering::lowerDYNAMIC_STACKALLOCImpl(SDValue Op,
   Chain = DAG.getCALLSEQ_START(Chain, 0, 0, dl);
 
   SDValue Size = Tmp2.getOperand(1);
-  SDValue SP = DAG.getCopyFromReg(Chain, dl, SPReg, VT);
-  Chain = SP.getValue(1);
-  MaybeAlign Alignment = cast<ConstantSDNode>(Tmp3)->getMaybeAlignValue();
+  SDValue BaseAddr = DAG.getCopyFromReg(Chain, dl, SPReg, VT);
+  Align Alignment = cast<ConstantSDNode>(Tmp3)->getAlignValue();
+
   const TargetFrameLowering *TFL = Subtarget->getFrameLowering();
   assert(TFL->getStackGrowthDirection() == TargetFrameLowering::StackGrowsUp &&
          "Stack grows upwards for AMDGPU");
 
+  Chain = BaseAddr.getValue(1);
+  Align StackAlign = TFL->getStackAlign();
+  if (Alignment > StackAlign) {
+    auto ScaledAlignment = (uint64_t)Alignment.value()
----------------
s-barannikov wrote:

`auto` is discouraged by the coding standard.

https://github.com/llvm/llvm-project/pull/119822