AMDGPU: Fix a SIAnnotateControlFlow issue when there are multiple backedges.

Summary:
At the exit of the loop, the compiler uses a register to remember and accumulate
the number of threads that have already exited. When all active threads exit the
loop, this register is used to restore the exec mask, and the execution continues
for the post loop code.

When there is a "continue" in the loop, the compiler made a mistake to reset the
register to 0 when the "continue" backedge is taken. This will result in some
threads not executing the post loop code as they are supposed to.

This patch fixed the issue.

Reviewers:
  nhaehnle, arsenm

Differential Revision:
  https://reviews.llvm.org/D59312

llvm-svn: 356298

1 files changed, 11 insertions, 2 deletions

diff --git a/llvm/lib/Target/AMDGPU/SIAnnotateControlFlow.cpp b/llvm/lib/Target/AMDGPU/SIAnnotateControlFlow.cpp
index 45642288948..a26c12bb934 100644
--- a/llvm/lib/Target/AMDGPU/SIAnnotateControlFlow.cpp
+++ b/llvm/lib/Target/AMDGPU/SIAnnotateControlFlow.cpp
@@ -264,8 +264,17 @@ void SIAnnotateControlFlow::handleLoop(BranchInst *Term) {
   Term->setCondition(BoolTrue);
   Value *Arg = handleLoopCondition(Cond, Broken, L, Term);
 
-  for (BasicBlock *Pred : predecessors(Target))
-    Broken->addIncoming(Pred == BB ? Arg : Int64Zero, Pred);
+  for (BasicBlock *Pred : predecessors(Target)) {
+    Value *PHIValue = Int64Zero;
+    if (Pred == BB) // Remember the value of the previous iteration.
+      PHIValue = Arg;
+    // If the backedge from Pred to Target could be executed before the exit
+    // of the loop at BB, it should not reset or change "Broken", which keeps
+    // track of the number of threads exited the loop at BB.
+    else if (L->contains(Pred) && DT->dominates(Pred, BB))
+      PHIValue = Broken;
+    Broken->addIncoming(PHIValue, Pred);
+  }
 
   Term->setCondition(CallInst::Create(Loop, Arg, "", Term));