[llvm] r274275 - AMDGPU: Add m0 vgpr load loop block as successor

[Matt Arsenault via llvm-commits]

Author: arsenm
Date: Thu Jun 30 15:49:28 2016
New Revision: 274275

URL: http://llvm.org/viewvc/llvm-project?rev=274275&view=rev
Log:
AMDGPU: Add m0 vgpr load loop block as successor

This shows up as a verifier error when I move this
earlier, not sure why it didn't before.

Modified:
    llvm/trunk/lib/Target/AMDGPU/SILowerControlFlow.cpp

Modified: llvm/trunk/lib/Target/AMDGPU/SILowerControlFlow.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/AMDGPU/SILowerControlFlow.cpp?rev=274275&r1=274274&r2=274275&view=diff
==============================================================================
--- llvm/trunk/lib/Target/AMDGPU/SILowerControlFlow.cpp (original)
+++ llvm/trunk/lib/Target/AMDGPU/SILowerControlFlow.cpp Thu Jun 30 15:49:28 2016
@@ -521,6 +521,7 @@ bool SILowerControlFlow::loadM0(MachineI
   // Move the rest of the block into a new block.
   RemainderBB->transferSuccessors(&MBB);
   RemainderBB->splice(RemainderBB->begin(), &MBB, I, MBB.end());
+  MBB.addSuccessor(LoopBB);
 
   emitLoadM0FromVGPRLoop(*LoopBB, DL, MovRel, *Idx, Offset);