From 7cc27649a68170c6da3cecbde94e8a14dd56346e Mon Sep 17 00:00:00 2001
From: Akira Hatanaka <ahatanaka@apple.com>
Date: Thu, 10 Jul 2014 18:00:53 +0000
Subject: [PATCH] [X86] Mark pseudo instruction TEST8ri_NOEREX as
 hasSIdeEffects=0.

Also, add a case clause in X86InstrInfo::shouldScheduleAdjacent to enable
macro-fusion.

<rdar://problem/15680770>

llvm-svn: 212747
---
 llvm/lib/Target/X86/X86InstrArithmetic.td |  6 ++++--
 llvm/lib/Target/X86/X86InstrInfo.cpp      |  1 +
 llvm/test/CodeGen/X86/testb-je-fusion.ll  | 20 ++++++++++++++++++++
 3 files changed, 25 insertions(+), 2 deletions(-)
 create mode 100644 llvm/test/CodeGen/X86/testb-je-fusion.ll

diff --git a/llvm/lib/Target/X86/X86InstrArithmetic.td b/llvm/lib/Target/X86/X86InstrArithmetic.td
index 368e14b..f2574cc 100644
--- a/llvm/lib/Target/X86/X86InstrArithmetic.td
+++ b/llvm/lib/Target/X86/X86InstrArithmetic.td
@@ -1278,8 +1278,10 @@ let isCompare = 1 in {
     def TEST64mi32 : BinOpMI_F<"test", Xi64, X86testpat, MRM0m, 0xF6>;
 
     // When testing the result of EXTRACT_SUBREG sub_8bit_hi, make sure the
-    // register class is constrained to GR8_NOREX.
-    let isPseudo = 1 in
+    // register class is constrained to GR8_NOREX. This pseudo is explicitly
+    // marked side-effect free, since it doesn't have an isel pattern like
+    // other test instructions. 
+    let isPseudo = 1, hasSideEffects = 0 in
     def TEST8ri_NOREX : I<0, Pseudo, (outs), (ins GR8_NOREX:$src, i8imm:$mask),
                           "", [], IIC_BIN_NONMEM>, Sched<[WriteALU]>;
   } // Defs = [EFLAGS]
diff --git a/llvm/lib/Target/X86/X86InstrInfo.cpp b/llvm/lib/Target/X86/X86InstrInfo.cpp
index 0797fc6..0d3afc4 100644
--- a/llvm/lib/Target/X86/X86InstrInfo.cpp
+++ b/llvm/lib/Target/X86/X86InstrInfo.cpp
@@ -5037,6 +5037,7 @@ bool X86InstrInfo::shouldScheduleAdjacent(MachineInstr* First,
   case X86::TEST16rm:
   case X86::TEST32rm:
   case X86::TEST64rm:
+  case X86::TEST8ri_NOREX:
   case X86::AND16i16:
   case X86::AND16ri:
   case X86::AND16ri8:
diff --git a/llvm/test/CodeGen/X86/testb-je-fusion.ll b/llvm/test/CodeGen/X86/testb-je-fusion.ll
new file mode 100644
index 0000000..9e946ae
--- /dev/null
+++ b/llvm/test/CodeGen/X86/testb-je-fusion.ll
@@ -0,0 +1,20 @@
+; RUN: llc < %s -march=x86-64 -mcpu=corei7-avx | FileCheck %s
+
+; testb should be scheduled right before je to enable macro-fusion.
+
+; CHECK: testb $2, %{{[abcd]}}h
+; CHECK-NEXT: je
+
+define i32 @check_flag(i32 %flags, ...) nounwind {
+entry:
+  %and = and i32 %flags, 512
+  %tobool = icmp eq i32 %and, 0
+  br i1 %tobool, label %if.end, label %if.then
+
+if.then:
+  br label %if.end
+
+if.end:
+  %hasflag = phi i32 [ 1, %if.then ], [ 0, %entry ]
+  ret i32 %hasflag
+}
-- 
2.7.4