[SelectionDAGISel] Don't merge input chains if it would put a token factor in the way of a glue. (#167805)

In the new test, we're trying to fold a load and a X86ISD::CALL. The
call has a CopyToReg glued to it. The load and the call have different
input chains so they need to be merged. This results in a TokenFactor
that gets put between the CopyToReg and the final CALLm instruction. The
DAG scheduler can't handle that.

The load here was created by legalization of the extract_element using a
stack temporary store and load. A normal IR load would be chained into
call sequence by SelectionDAGBuilder. This would usually have the load
chained in before the CopyToReg. The store/load created by legalization
don't get chained into the rest of the DAG.

Fixes #63790

Author: topperc

llvm/lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp

@@ -2783,7 +2783,7 @@ void SelectionDAGISel::UpdateChains(
 /// be used as the input node chain for the generated nodes.
 static SDValue
 HandleMergeInputChains(const SmallVectorImpl<SDNode *> &ChainNodesMatched,
-                       SelectionDAG *CurDAG) {
+                       SDValue InputGlue, SelectionDAG *CurDAG) {
 
   SmallPtrSet<const SDNode *, 16> Visited;
   SmallVector<const SDNode *, 8> Worklist;
@@ -2826,8 +2826,16 @@ HandleMergeInputChains(const SmallVectorImpl<SDNode *> &ChainNodesMatched,
   // node that is both the predecessor and successor of the
   // to-be-merged nodes. Fail.
   Visited.clear();
-  for (SDValue V : InputChains)
+  for (SDValue V : InputChains) {
+    // If we need to create a TokenFactor, and any of the input chain nodes will
+    // also be glued to the output, we cannot merge the chains. The TokenFactor
+    // would prevent the glue from being honored.
+    if (InputChains.size() != 1 &&
+        V->getValueType(V->getNumValues() - 1) == MVT::Glue &&
+        InputGlue.getNode() == V.getNode())
+      return SDValue();
     Worklist.push_back(V.getNode());
+  }
 
   for (auto *N : ChainNodesMatched)
     if (SDNode::hasPredecessorHelper(N, Visited, Worklist, Max, true))
@@ -3989,7 +3997,7 @@ void SelectionDAGISel::SelectCodeCommon(SDNode *NodeToMatch,
       }
 
       // Merge the input chains if they are not intra-pattern references.
-      InputChain = HandleMergeInputChains(ChainNodesMatched, CurDAG);
+      InputChain = HandleMergeInputChains(ChainNodesMatched, InputGlue, CurDAG);
 
       if (!InputChain.getNode())
         break;  // Failed to merge.
@@ -4033,7 +4041,7 @@ void SelectionDAGISel::SelectCodeCommon(SDNode *NodeToMatch,
         break;
 
       // Merge the input chains if they are not intra-pattern references.
-      InputChain = HandleMergeInputChains(ChainNodesMatched, CurDAG);
+      InputChain = HandleMergeInputChains(ChainNodesMatched, InputGlue, CurDAG);
 
       if (!InputChain.getNode())
         break;  // Failed to merge.

llvm/test/CodeGen/X86/pr63790.ll

@@ -0,0 +1,26 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 6
+; RUN: llc < %s -mtriple=x86_64 | FileCheck %s
+
+define void @f(ptr %0, i64 %1) {
+; CHECK-LABEL: f:
+; CHECK:       # %bb.0: # %BB
+; CHECK-NEXT:    subq $40, %rsp
+; CHECK-NEXT:    .cfi_def_cfa_offset 48
+; CHECK-NEXT:    andl $1, %esi
+; CHECK-NEXT:    movaps (%rdi), %xmm0
+; CHECK-NEXT:    movaps %xmm0, (%rsp) # 16-byte Spill
+; CHECK-NEXT:    movaps %xmm0, {{[0-9]+}}(%rsp)
+; CHECK-NEXT:    movl $42, %edi
+; CHECK-NEXT:    callq *16(%rsp,%rsi,8)
+; CHECK-NEXT:    movaps (%rsp), %xmm0 # 16-byte Reload
+; CHECK-NEXT:    movaps %xmm0, (%rax)
+; CHECK-NEXT:    addq $40, %rsp
+; CHECK-NEXT:    .cfi_def_cfa_offset 8
+; CHECK-NEXT:    retq
+BB:
+  %fps = load <2 x ptr>, ptr %0
+  %fp = extractelement <2 x ptr> %fps, i64 %1
+  %p = call ptr %fp(i32 42)
+  store <2 x ptr> %fps, ptr %p
+  ret void
+}