From ef4dbb46daf007bd1357ef3bc1ae95a40956a360 Mon Sep 17 00:00:00 2001
From: Changpeng Fang <changpeng.fang@gmail.com>
Date: Thu, 3 Aug 2017 16:37:02 +0000
Subject: [PATCH] AMDGPU/SI: Don't fix a PHI under uniform branch in
 SIFixSGPRCopies only when sources and destination are all sgprs

Summary:
  If a PHI has at lease one VGPR operand, we have to fix the PHI
in SIFixSGPRCopies.

Reviewer:
  Matt

Differential Revision:
  http://reviews.llvm.org/D34727

llvm-svn: 309959
---
 llvm/lib/Target/AMDGPU/SIFixSGPRCopies.cpp |  6 +++---
 llvm/test/CodeGen/AMDGPU/uniform-cfg.ll    | 22 ++++++++++++++++++++++
 2 files changed, 25 insertions(+), 3 deletions(-)

diff --git a/llvm/lib/Target/AMDGPU/SIFixSGPRCopies.cpp b/llvm/lib/Target/AMDGPU/SIFixSGPRCopies.cpp
index 0a795c99f94e..32e83cb385f0 100644
--- a/llvm/lib/Target/AMDGPU/SIFixSGPRCopies.cpp
+++ b/llvm/lib/Target/AMDGPU/SIFixSGPRCopies.cpp
@@ -604,7 +604,8 @@ bool SIFixSGPRCopies::runOnMachineFunction(MachineFunction &MF) {
 
         // We don't need to fix the PHI if the common dominator of the
         // two incoming blocks terminates with a uniform branch.
-        if (MI.getNumExplicitOperands() == 5) {
+        bool HasVGPROperand = phiHasVGPROperands(MI, MRI, TRI, TII);
+        if (MI.getNumExplicitOperands() == 5 && !HasVGPROperand) {
           MachineBasicBlock *MBB0 = MI.getOperand(2).getMBB();
           MachineBasicBlock *MBB1 = MI.getOperand(4).getMBB();
 
@@ -649,8 +650,7 @@ bool SIFixSGPRCopies::runOnMachineFunction(MachineFunction &MF) {
         // is no chance for values to be over-written.
 
         SmallSet<unsigned, 8> Visited;
-        if (phiHasVGPROperands(MI, MRI, TRI, TII) ||
-            !phiHasBreakDef(MI, MRI, Visited)) {
+        if (HasVGPROperand || !phiHasBreakDef(MI, MRI, Visited)) {
           DEBUG(dbgs() << "Fixing PHI: " << MI);
           TII->moveToVALU(MI);
         }
diff --git a/llvm/test/CodeGen/AMDGPU/uniform-cfg.ll b/llvm/test/CodeGen/AMDGPU/uniform-cfg.ll
index 9416336a44ee..3c25dc7f0e79 100644
--- a/llvm/test/CodeGen/AMDGPU/uniform-cfg.ll
+++ b/llvm/test/CodeGen/AMDGPU/uniform-cfg.ll
@@ -557,6 +557,28 @@ done:
   ret void
 }
 
+; GCN-LABEL: {{^}}move_to_valu_vgpr_operand_phi:
+; GCN: v_add_i32_e32
+; GCN: ds_write_b32
+define void @move_to_valu_vgpr_operand_phi(i32 addrspace(3)* %out) {
+bb0:
+  br label %bb1
+
+bb1:                                              ; preds = %bb3, %bb0
+  %tmp0 = phi i32 [ 8, %bb0 ], [ %tmp4, %bb3 ]
+  %tmp1 = add nsw i32 %tmp0, -1
+  %tmp2 = getelementptr inbounds i32, i32 addrspace(3)* %out, i32 %tmp1
+  br i1 undef, label %bb2, label %bb3
+
+bb2:                                              ; preds = %bb1
+  store volatile i32 1, i32 addrspace(3)* %tmp2, align 4
+  br label %bb3
+
+bb3:                                              ; preds = %bb2, %bb1
+  %tmp4 = add nsw i32 %tmp0, 2
+  br label %bb1
+}
+
 declare i32 @llvm.amdgcn.workitem.id.x() #0
 
 attributes #0 = { nounwind readnone }