Add a late optimization to BranchFolding that hoist common instruction sequences

at the start of basic blocks to their common predecessor. It's actually quite
common (e.g. about 50 times in JM/lencod) and has shown to be a nice code size
benefit. e.g.

        pushq   %rax
        testl   %edi, %edi
        jne     LBB0_2
## BB#1:
        xorb    %al, %al
        popq    %rdx
        ret
LBB0_2:
        xorb    %al, %al
        callq   _foo
        popq    %rdx
        ret

=>

        pushq   %rax
        xorb    %al, %al
        testl   %edi, %edi
        je      LBB0_2
## BB#1:
        callq   _foo
LBB0_2:
        popq    %rdx
        ret

rdar://9145558

llvm-svn: 131172

1 files changed, 2 insertions, 2 deletions

diff --git a/llvm/lib/CodeGen/IfConversion.cpp b/llvm/lib/CodeGen/IfConversion.cpp
index 790200b8df5..8b2c9816164 100644
--- a/llvm/lib/CodeGen/IfConversion.cpp
+++ b/llvm/lib/CodeGen/IfConversion.cpp
@@ -265,7 +265,7 @@ bool IfConverter::runOnMachineFunction(MachineFunction &MF) {
   if (!TII) return false;
 
   // Tail merge tend to expose more if-conversion opportunities.
-  BranchFolder BF(true);
+  BranchFolder BF(true, false);
   bool BFChange = BF.OptimizeFunction(MF, TII,
                                    MF.getTarget().getRegisterInfo(),
                                    getAnalysisIfAvailable<MachineModuleInfo>());
@@ -399,7 +399,7 @@ bool IfConverter::runOnMachineFunction(MachineFunction &MF) {
   BBAnalysis.clear();
 
   if (MadeChange && IfCvtBranchFold) {
-    BranchFolder BF(false);
+    BranchFolder BF(false, false);
     BF.OptimizeFunction(MF, TII,
                         MF.getTarget().getRegisterInfo(),
                         getAnalysisIfAvailable<MachineModuleInfo>());