[MachinePipeliner] Fix order for nodes with Anti dependence in same cycle

Summary:
Problem exposed in PowerPC functional testing.

We did not consider Anti dependence for nodes in same cycle,
so we may end up generating bad machine code.
eg: the reduced test won't verify.

*** Bad machine code: Using an undefined physical register ***
- function:    lame_encode_buffer_interleaved
- basic block: %bb.4  (0x4bde4e12928)
- instruction: %29:gprc = ADDZE %27:gprc, implicit-def dead $carry, implicit $carry
- operand 3:   implicit $carry

Reviewers: bcahoon, kparzysz, hfinkel

Subscribers: MaskRay, wuzish, nemanjai, hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D64192

llvm-svn: 365859

1 files changed, 8 insertions, 0 deletions

diff --git a/llvm/lib/CodeGen/MachinePipeliner.cpp b/llvm/lib/CodeGen/MachinePipeliner.cpp
index 9c0c5cc5c70..54df522d371 100644
--- a/llvm/lib/CodeGen/MachinePipeliner.cpp
+++ b/llvm/lib/CodeGen/MachinePipeliner.cpp
@@ -3559,6 +3559,14 @@ void SMSchedule::orderDependence(SwingSchedulerDAG *SSD, SUnit *SU,
         if (Pos < MoveUse)
           MoveUse = Pos;
       }
+      // We did not handle HW dependences in previous for loop,
+      // and we normally set Latency = 0 for Anti deps,
+      // so may have nodes in same cycle with Anti denpendent on HW regs.
+      else if (S.getKind() == SDep::Anti && stageScheduled(*I) == StageInst1) {
+        OrderBeforeUse = true;
+        if ((MoveUse == 0) || (Pos < MoveUse))
+          MoveUse = Pos;
+      }
     }
     for (auto &P : SU->Preds) {
       if (P.getSUnit() != *I)