AMDGPU: Use correct chain when emitting error on a call

[arsenm]

Return the input chain at the callsite, not the entry node
chain. Presumably this could cause issues somewhere.

[arsenm]

• DAG: Check if deoptimize call is available before emitting it #170940
• AMDGPU: Avoid crashing on statepoint-like pseudoinstructions #170657
• AMDGPU: Create a dummy call sequence when emitting call error #170656
• AMDGPU: Use correct chain when emitting error on a call #170645 👈 (View in Graphite)
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[llvmbot]

@llvm/pr-subscribers-backend-amdgpu

Author: Matt Arsenault (arsenm)

Changes

Return the input chain at the callsite, not the entry node
chain. Presumably this could cause issues somewhere.

---

Full diff: https://github.com/llvm/llvm-project/pull/170645.diff

1 Files Affected:

• (modified) llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp (+1-1)

diff --git a/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp b/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp
index 5be5c66ba17dc..ba80c0695115c 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUISelLowering.cpp
@@ -1412,7 +1412,7 @@ SDValue AMDGPUTargetLowering::lowerUnhandledCall(CallLoweringInfo &CLI,
       InVals.push_back(DAG.getPOISON(Arg.VT));
   }
 
-  return DAG.getEntryNode();
+  return CLI.Chain;
 }
 
 SDValue AMDGPUTargetLowering::LowerCall(CallLoweringInfo &CLI,

[ro-i]

seems to make sense to me to not throw away everything the chain did up to that point