diff --git a/clang/include/clang/Basic/BuiltinsPPC.def b/clang/include/clang/Basic/BuiltinsPPC.def
--- a/clang/include/clang/Basic/BuiltinsPPC.def
+++ b/clang/include/clang/Basic/BuiltinsPPC.def
@@ -57,8 +57,11 @@
 BUILTIN(__builtin_ppc_fetch_and_swaplp, "ULiULiD*ULi", "")
 BUILTIN(__builtin_ppc_ldarx, "LiLiD*", "")
 BUILTIN(__builtin_ppc_lwarx, "iiD*", "")
+BUILTIN(__builtin_ppc_lharx, "isD*", "")
+BUILTIN(__builtin_ppc_lbarx, "UiUcD*", "")
 BUILTIN(__builtin_ppc_stdcx, "iLiD*Li", "")
 BUILTIN(__builtin_ppc_stwcx, "iiD*i", "")
+BUILTIN(__builtin_ppc_sthcx, "isD*s", "")
 BUILTIN(__builtin_ppc_tdw, "vLLiLLiIUi", "")
 BUILTIN(__builtin_ppc_tw, "viiIUi", "")
 BUILTIN(__builtin_ppc_trap, "vi", "")
@@ -71,6 +74,11 @@
 BUILTIN(__builtin_ppc_fctiwz, "dd", "")
 BUILTIN(__builtin_ppc_fctudz, "dd", "")
 BUILTIN(__builtin_ppc_fctuwz, "dd", "")
+BUILTIN(__builtin_ppc_dcbtstt, "vv*", "")
+BUILTIN(__builtin_ppc_dcbtt, "vv*", "")
+BUILTIN(__builtin_ppc_mftbu, "Ui","")
+BUILTIN(__builtin_ppc_mfmsr, "Ui", "")
+BUILTIN(__builtin_ppc_stfiw, "viC*d", "")
 
 BUILTIN(__builtin_ppc_get_timebase, "ULLi", "n")
 
diff --git a/clang/lib/Basic/Targets/PPC.cpp b/clang/lib/Basic/Targets/PPC.cpp
--- a/clang/lib/Basic/Targets/PPC.cpp
+++ b/clang/lib/Basic/Targets/PPC.cpp
@@ -116,8 +116,12 @@
   Builder.defineMacro("__fetch_and_swaplp", "__builtin_ppc_fetch_and_swaplp");
   Builder.defineMacro("__ldarx", "__builtin_ppc_ldarx");
   Builder.defineMacro("__lwarx", "__builtin_ppc_lwarx");
+  Builder.defineMacro("__lharx", "__builtin_ppc_lharx");
+  Builder.defineMacro("__lbarx", "__builtin_ppc_lbarx");
+  Builder.defineMacro("__stfiw", "__builtin_ppc_stfiw");
   Builder.defineMacro("__stdcx", "__builtin_ppc_stdcx");
   Builder.defineMacro("__stwcx", "__builtin_ppc_stwcx");
+  Builder.defineMacro("__sthcx", "__builtin_ppc_sthcx");
   Builder.defineMacro("__tdw", "__builtin_ppc_tdw");
   Builder.defineMacro("__tw", "__builtin_ppc_tw");
   Builder.defineMacro("__trap", "__builtin_ppc_trap");
@@ -130,6 +134,10 @@
   Builder.defineMacro("__fctiwz", "__builtin_ppc_fctiwz");
   Builder.defineMacro("__fctudz", "__builtin_ppc_fctudz");
   Builder.defineMacro("__fctuwz", "__builtin_ppc_fctuwz");
+  Builder.defineMacro("__dcbtstt", "__builtin_ppc_dcbtstt");
+  Builder.defineMacro("__dcbtt", "__builtin_ppc_dcbtt");
+  Builder.defineMacro("__mftbu", "__builtin_ppc_mftbu");
+  Builder.defineMacro("__mfmsr", "__builtin_ppc_mfmsr");
 }
 
 /// PPCTargetInfo::getTargetDefines - Return a set of the PowerPC-specific
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -1012,6 +1012,14 @@
     AsmOS << "lwarx ";
     RetType = CGF.Int32Ty;
     break;
+  case clang::PPC::BI__builtin_ppc_lharx:
+    AsmOS << "lharx ";
+    RetType = CGF.Int16Ty;
+    break;
+  case clang::PPC::BI__builtin_ppc_lbarx:
+    AsmOS << "lbarx ";
+    RetType = CGF.Int8Ty;
+    break;
   default:
     llvm_unreachable("Expected only PowerPC load reserve intrinsics");
   }
@@ -15464,6 +15472,13 @@
     return Builder.CreateExtractElement(Unpacked, Index);
   }
 
+  case PPC::BI__builtin_ppc_sthcx: {
+    llvm::Function *F = CGM.getIntrinsic(Intrinsic::ppc_sthcx);
+    Ops[0] = Builder.CreateBitCast(Ops[0], Int8PtrTy);
+    Ops[1] = Builder.CreateSExt(Ops[1], Int32Ty);
+    return Builder.CreateCall(F, Ops);
+  }
+
   // The PPC MMA builtins take a pointer to a __vector_quad as an argument.
   // Some of the MMA instructions accumulate their result into an existing
   // accumulator whereas the others generate a new accumulator. So we need to
@@ -15574,6 +15589,8 @@
   }
   case PPC::BI__builtin_ppc_ldarx:
   case PPC::BI__builtin_ppc_lwarx:
+  case PPC::BI__builtin_ppc_lharx:
+  case PPC::BI__builtin_ppc_lbarx:
     return emitPPCLoadReserveIntrinsic(*this, BuiltinID, E);
   }
 }
diff --git a/clang/lib/Sema/SemaChecking.cpp b/clang/lib/Sema/SemaChecking.cpp
--- a/clang/lib/Sema/SemaChecking.cpp
+++ b/clang/lib/Sema/SemaChecking.cpp
@@ -3360,6 +3360,11 @@
   case PPC::BI__builtin_ppc_tw:
   case PPC::BI__builtin_ppc_tdw:
     return SemaBuiltinConstantArgRange(TheCall, 2, 0, 31);
+  case PPC::BI__builtin_ppc_sthcx:
+  case PPC::BI__builtin_ppc_lharx:
+  case PPC::BI__builtin_ppc_lbarx:
+    return SemaFeatureCheck(*this, TheCall, "isa-v207-instructions",
+                            diag::err_ppc_builtin_only_on_arch, "8");
 #define CUSTOM_BUILTIN(Name, Intr, Types, Acc) \
   case PPC::BI__builtin_##Name: \
     return SemaBuiltinPPCMMACall(TheCall, Types);
diff --git a/clang/test/CodeGen/builtins-ppc-xlcompat-LoadReseve-StoreCond.c b/clang/test/CodeGen/builtins-ppc-xlcompat-LoadReseve-StoreCond.c
--- a/clang/test/CodeGen/builtins-ppc-xlcompat-LoadReseve-StoreCond.c
+++ b/clang/test/CodeGen/builtins-ppc-xlcompat-LoadReseve-StoreCond.c
@@ -1,20 +1,47 @@
-// RUN: %clang_cc1 -O2 -triple=powerpc-unknown-aix -emit-llvm %s -o - | \
-// RUN: FileCheck %s
-// RUN: %clang_cc1 -O2 -triple=powerpc64-unknown-aix -emit-llvm %s -o - | \
-// RUN: FileCheck %s
-// RUN: %clang_cc1 -O2 -triple=powerpc64le-unknown-unknown -emit-llvm %s \
-// RUN:  -o - | FileCheck %s
-// RUN: %clang_cc1 -O2 -triple=powerpc64-unknown-unknown -emit-llvm %s \
-// RUN:  -o - | FileCheck %s
+// RUN: %clang_cc1 -O2 -target-cpu pwr8 -triple=powerpc-unknown-aix \
+// RUN:  -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -O2 -target-cpu pwr8 -triple=powerpc64-unknown-aix \
+// RUN:  -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -O2 -target-cpu pwr8 -triple=powerpc64le-unknown-unknown \
+// RUN:  -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -O2 -target-cpu pwr8 -triple=powerpc64-unknown-unknown \
+// RUN:  -emit-llvm %s -o - | FileCheck %s
+// RAUN: not %clang_cc1 -O2 -target-cpu pwr7 -triple=powerpc-unknown-aix \
+// RAUN:  -emit-llvm %s -o - 2>&1 | FileCheck %s \
+// RAUN:  --check-prefix=CHECK-NON-PWR8-ERR
 
 int test_lwarx(volatile int* a) {
-  // CHECK: @test_lwarx
+  // CHECK-LABEL: @test_lwarx
   // CHECK: %0 = tail call i32 asm sideeffect "lwarx $0, ${1:y}", "=r,*Z,~{memory}"(i32* %a)
   return __lwarx(a);
 }
+
+short test_lharx(volatile short *a) {
+  // CHECK-LABEL: @test_lharx
+  // CHECK: %0 = tail call i16 asm sideeffect "lharx $0, ${1:y}", "=r,*Z,~{memory}"(i16* %a)
+  // CHECK-NON-PWR8-ERR:  error: this builtin is only valid on POWER8 or later CPUs
+  return __lharx(a);
+}
+
+char test_lbarx(volatile unsigned char *a) {
+  // CHECK-LABEL: @test_lbarx
+  // CHECK: %0 = tail call i8 asm sideeffect "lbarx $0, ${1:y}", "=r,*Z,~{memory}"(i8* %a)
+  // CHECK-NON-PWR8-ERR:  error: this builtin is only valid on POWER8 or later CPUs
+  return __lbarx(a);
+}
+
 int test_stwcx(volatile int* a, int val) {
-  // CHECK: @test_stwcx
+  // CHECK-LABEL: @test_stwcx
   // CHECK: %0 = bitcast i32* %a to i8*
   // CHECK: %1 = tail call i32 @llvm.ppc.stwcx(i8* %0, i32 %val)
   return __stwcx(a, val);
 }
+
+int test_sthcx(volatile short *a, short val) {
+  // CHECK-LABEL: @test_sthcx
+  // CHECK: %0 = bitcast i16* %a to i8*
+  // CHECK: %1 = sext i16 %val to i32
+  // CHECK: %2 = tail call i32 @llvm.ppc.sthcx(i8* %0, i32 %1)
+  // CHECK-NON-PWR8-ERR:  error: this builtin is only valid on POWER8 or later CPUs
+  return __sthcx(a, val);
+}
diff --git a/clang/test/CodeGen/builtins-ppc-xlcompat-move-tofrom-regs.c b/clang/test/CodeGen/builtins-ppc-xlcompat-move-tofrom-regs.c
new file mode 100644
--- /dev/null
+++ b/clang/test/CodeGen/builtins-ppc-xlcompat-move-tofrom-regs.c
@@ -0,0 +1,20 @@
+// RUN: %clang_cc1 -O2 -triple powerpc64-unknown-unknown \
+// RUN:   -emit-llvm %s -o - -target-cpu pwr7 | FileCheck %s
+// RUN: %clang_cc1 -O2 -triple powerpc64le-unknown-unknown \
+// RUN:   -emit-llvm %s -o - -target-cpu pwr8 | FileCheck %s
+// RUN: %clang_cc1 -O2 -triple powerpc-unknown-aix \
+// RUN:   -emit-llvm %s -o - -target-cpu pwr7 | FileCheck %s
+// RUN: %clang_cc1 -O2 -triple powerpc64-unknown-aix \
+// RUN:   -emit-llvm %s -o - -target-cpu pwr7 | FileCheck %s
+
+unsigned int test_mftbu(void) {
+  // CHECK-LABEL: @test_mftbu
+  // CHECK: %0 = tail call i32 @llvm.ppc.mftbu()
+  return __mftbu();
+}
+
+unsigned long test_mfmsr(void) {
+  // CHECK-LABEL: @test_mfmsr
+  // CHECK: %0 = tail call i32 @llvm.ppc.mfmsr()
+  return __mfmsr();
+}
diff --git a/clang/test/CodeGen/builtins-ppc-xlcompat-prefetch.c b/clang/test/CodeGen/builtins-ppc-xlcompat-prefetch.c
new file mode 100644
--- /dev/null
+++ b/clang/test/CodeGen/builtins-ppc-xlcompat-prefetch.c
@@ -0,0 +1,26 @@
+// RUN: %clang_cc1 -O2 -triple powerpc64-unknown-unknown \
+// RUN:   -emit-llvm %s -o - -target-cpu pwr7 | FileCheck %s
+// RUN: %clang_cc1 -O2 -triple powerpc64le-unknown-unknown \
+// RUN:   -emit-llvm %s -o - -target-cpu pwr8 | FileCheck %s
+// RUN: %clang_cc1 -O2 -triple powerpc-unknown-aix \
+// RUN:   -emit-llvm %s -o - -target-cpu pwr7 | FileCheck %s
+// RUN: %clang_cc1 -O2 -triple powerpc64-unknown-aix \
+// RUN:   -emit-llvm %s -o - -target-cpu pwr7 | FileCheck %s
+
+extern void *vpa;
+
+void test_dcbtstt(void) {
+  // CHECK-LABEL: @test_dcbtstt
+  // CHECK: %0 = load i8*, i8** @vpa
+  // CHECK: tail call void @llvm.ppc.dcbtstt(i8* %0)
+  // CHECK: ret void
+  __dcbtstt(vpa);
+}
+
+void test_dcbtt(void) {
+  // CHECK-LABEL: @test_dcbt
+  // CHECK: %0 = load i8*, i8** @vpa
+  // CHECK: tail call void @llvm.ppc.dcbtt(i8* %0)
+  // CHECK: ret void
+  __dcbtt(vpa);
+}
diff --git a/clang/test/CodeGen/builtins-ppc-xlcompat-stfiw.c b/clang/test/CodeGen/builtins-ppc-xlcompat-stfiw.c
new file mode 100644
--- /dev/null
+++ b/clang/test/CodeGen/builtins-ppc-xlcompat-stfiw.c
@@ -0,0 +1,23 @@
+// RUN: %clang_cc1 -O2 -triple powerpc64-unknown-unknown \
+// RUN:   -emit-llvm %s -o - -target-cpu pwr7 | FileCheck %s
+// RUN: %clang_cc1 -O2 -triple powerpc64le-unknown-unknown \
+// RUN:   -emit-llvm %s -o - -target-cpu pwr8 | FileCheck %s
+// RUN: %clang_cc1 -O2 -triple powerpc-unknown-aix \
+// RUN:   -emit-llvm %s -o - -target-cpu pwr8 | FileCheck %s
+// RUN: %clang_cc1 -O2 -triple powerpc64-unknown-aix \
+// RUN:   -emit-llvm %s -o - -target-cpu pwr8 | FileCheck %s
+
+extern const int *cia;
+extern double da;
+
+void test_stfiw() {
+  // CHECK-LABEL: test_stfiw
+  // CHECK: void @llvm.ppc.stfiw(i8* %0, double %1)
+  __builtin_ppc_stfiw(cia, da);
+}
+
+void test_xl_stfiw() {
+  // CHECK-LABEL: test_xl_stfiw
+  // CHECK: void @llvm.ppc.stfiw(i8* %0, double %1)
+  __stfiw(cia, da);
+}
diff --git a/llvm/include/llvm/IR/IntrinsicsPowerPC.td b/llvm/include/llvm/IR/IntrinsicsPowerPC.td
--- a/llvm/include/llvm/IR/IntrinsicsPowerPC.td
+++ b/llvm/include/llvm/IR/IntrinsicsPowerPC.td
@@ -1565,5 +1565,18 @@
   def int_ppc_stwcx : GCCBuiltin<"__builtin_ppc_stwcx">,
                       Intrinsic<[llvm_i32_ty], [llvm_ptr_ty, llvm_i32_ty],
                                 [IntrWriteMem]>;
+  def int_ppc_sthcx
+      : Intrinsic<[llvm_i32_ty], [ llvm_ptr_ty, llvm_i32_ty ], [IntrWriteMem]>;
+  def int_ppc_dcbtstt : GCCBuiltin<"__builtin_ppc_dcbtstt">,
+                        Intrinsic<[], [llvm_ptr_ty], [IntrWriteMem]>;
+  def int_ppc_dcbtt : GCCBuiltin<"__builtin_ppc_dcbtt">,
+                      Intrinsic<[], [llvm_ptr_ty], [IntrWriteMem]>;
+  def int_ppc_mftbu : GCCBuiltin<"__builtin_ppc_mftbu">,
+                      Intrinsic<[llvm_i32_ty], [], [IntrNoMem]>;
+  def int_ppc_mfmsr : GCCBuiltin<"__builtin_ppc_mfmsr">,
+                      Intrinsic<[llvm_i32_ty], [], [IntrNoMem]>;
+  def int_ppc_stfiw : GCCBuiltin<"__builtin_ppc_stfiw">,
+                      Intrinsic<[], [llvm_ptr_ty, llvm_double_ty],
+                                [IntrWriteMem]>;
 }
 
diff --git a/llvm/lib/Target/PowerPC/PPC.td b/llvm/lib/Target/PowerPC/PPC.td
--- a/llvm/lib/Target/PowerPC/PPC.td
+++ b/llvm/lib/Target/PowerPC/PPC.td
@@ -331,7 +331,8 @@
      FeatureDirectMove,
      FeatureICBT,
      FeaturePartwordAtomic,
-     FeaturePredictableSelectIsExpensive
+     FeaturePredictableSelectIsExpensive,
+     FeatureISA2_07
     ];
 
   list<SubtargetFeature> P8SpecificFeatures = [FeatureAddiLoadFusion,
diff --git a/llvm/lib/Target/PowerPC/PPCInstrInfo.td b/llvm/lib/Target/PowerPC/PPCInstrInfo.td
--- a/llvm/lib/Target/PowerPC/PPCInstrInfo.td
+++ b/llvm/lib/Target/PowerPC/PPCInstrInfo.td
@@ -5436,3 +5436,18 @@
         (XSCVDPUXDS $A)>;
 def : Pat<(int_ppc_fctuwz f64:$A),
         (XSCVDPUXWS $A)>;
+
+def : Pat<(int_ppc_mfmsr), (MFMSR)>;
+def : Pat<(int_ppc_mftbu), (MFTB 269)>;
+
+let Predicates = [IsISA2_07] in {
+  def : Pat<(int_ppc_sthcx ForceXForm:$dst, gprc:$A),
+          (STHCX gprc:$A, ForceXForm:$dst)>;
+}
+def : Pat<(int_ppc_dcbtstt ForceXForm:$dst),
+          (DCBTST 16, ForceXForm:$dst)>;
+def : Pat<(int_ppc_dcbtt ForceXForm:$dst),
+          (DCBT 16, ForceXForm:$dst)>;
+
+def : Pat<(int_ppc_stfiw ForceXForm:$dst, f64:$XT),
+          (STFIWX f64:$XT, ForceXForm:$dst)>;
diff --git a/llvm/lib/Target/PowerPC/PPCInstrVSX.td b/llvm/lib/Target/PowerPC/PPCInstrVSX.td
--- a/llvm/lib/Target/PowerPC/PPCInstrVSX.td
+++ b/llvm/lib/Target/PowerPC/PPCInstrVSX.td
@@ -4005,6 +4005,9 @@
             (f64 (PPCcv_fp_to_uint_in_vsr f64:$src)), ForceXForm:$dst, 1),
           (STXSIBX (XSCVDPUXWS f64:$src), ForceXForm:$dst)>;
 
+def : Pat<(int_ppc_stfiw ForceXForm:$dst, f64:$XT),
+          (STXSIWX f64:$XT, ForceXForm:$dst)>;
+
 // Round & Convert QP -> DP/SP
 def : Pat<(f64 (any_fpround f128:$src)), (f64 (XSCVQPDP $src))>;
 def : Pat<(f32 (any_fpround f128:$src)), (f32 (XSRSP (XSCVQPDPO $src)))>;
diff --git a/llvm/test/CodeGen/PowerPC/builtins-ppc-xlcompat-LoadReserve-StoreCond.ll b/llvm/test/CodeGen/PowerPC/builtins-ppc-xlcompat-LoadReserve-StoreCond.ll
--- a/llvm/test/CodeGen/PowerPC/builtins-ppc-xlcompat-LoadReserve-StoreCond.ll
+++ b/llvm/test/CodeGen/PowerPC/builtins-ppc-xlcompat-LoadReserve-StoreCond.ll
@@ -50,3 +50,69 @@
   %1 = tail call i32 @llvm.ppc.stwcx(i8* %0, i32 %b)
   ret i32 %1
 }
+
+declare i32 @llvm.ppc.sthcx(i8*, i32)
+define dso_local signext i32 @test_sthcx(i16* %a, i16 signext %val) {
+; CHECK-64-LABEL: test_sthcx:
+; CHECK-64:       # %bb.0: # %entry
+; CHECK-64-NEXT:    sthcx. 4, 0, 3
+; CHECK-64-NEXT:    mfocrf 3, 128
+; CHECK-64-NEXT:    srwi 3, 3, 28
+; CHECK-64-NEXT:    extsw 3, 3
+; CHECK-64-NEXT:    blr
+;
+; CHECK-32-LABEL: test_sthcx:
+; CHECK-32:       # %bb.0: # %entry
+; CHECK-32-NEXT:    sthcx. 4, 0, 3
+; CHECK-32-NEXT:    mfocrf 3, 128
+; CHECK-32-NEXT:    srwi 3, 3, 28
+; CHECK-32-NEXT:    blr
+entry:
+  %0 = bitcast i16* %a to i8*
+  %1 = sext i16 %val to i32
+  %2 = tail call i32 @llvm.ppc.sthcx(i8* %0, i32 %1)
+  ret i32 %2
+}
+
+define dso_local signext i16 @test_lharx(i16* %a) {
+; CHECK-64-LABEL: test_lharx:
+; CHECK-64:       # %bb.0: # %entry
+; CHECK-64-NEXT:    #APP
+; CHECK-64-NEXT:    lharx 3, 0, 3
+; CHECK-64-NEXT:    #NO_APP
+; CHECK-64-NEXT:    extsh 3, 3
+; CHECK-64-NEXT:    blr
+;
+; CHECK-32-LABEL: test_lharx:
+; CHECK-32:       # %bb.0: # %entry
+; CHECK-32-NEXT:    #APP
+; CHECK-32-NEXT:    lharx 3, 0, 3
+; CHECK-32-NEXT:    #NO_APP
+; CHECK-32-NEXT:    extsh 3, 3
+; CHECK-32-NEXT:    blr
+entry:
+  %0 = tail call i16 asm sideeffect "lharx $0, ${1:y}", "=r,*Z,~{memory}"(i16* %a)
+  ret i16 %0
+}
+
+; Function Attrs: nounwind uwtable
+define dso_local zeroext i8 @test_lbarx(i8* %a) {
+; CHECK-64-LABEL: test_lbarx:
+; CHECK-64:       # %bb.0: # %entry
+; CHECK-64-NEXT:    #APP
+; CHECK-64-NEXT:    lbarx 3, 0, 3
+; CHECK-64-NEXT:    #NO_APP
+; CHECK-64-NEXT:    clrldi 3, 3, 56
+; CHECK-64-NEXT:    blr
+;
+; CHECK-32-LABEL: test_lbarx:
+; CHECK-32:       # %bb.0: # %entry
+; CHECK-32-NEXT:    #APP
+; CHECK-32-NEXT:    lbarx 3, 0, 3
+; CHECK-32-NEXT:    #NO_APP
+; CHECK-32-NEXT:    clrlwi 3, 3, 24
+; CHECK-32-NEXT:    blr
+entry:
+  %0 = tail call i8 asm sideeffect "lbarx $0, ${1:y}", "=r,*Z,~{memory}"(i8* %a)
+  ret i8 %0
+}
diff --git a/llvm/test/CodeGen/PowerPC/builtins-ppc-xlcompat-stfiw.ll b/llvm/test/CodeGen/PowerPC/builtins-ppc-xlcompat-stfiw.ll
new file mode 100644
--- /dev/null
+++ b/llvm/test/CodeGen/PowerPC/builtins-ppc-xlcompat-stfiw.ll
@@ -0,0 +1,56 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu \
+; RUN:   -mcpu=pwr8 < %s | FileCheck %s
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu \
+; RUN:   -mcpu=pwr9 -mattr -vsx < %s | FileCheck %s --check-prefix=CHECK-NO-VSX
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu \
+; RUN:   -mcpu=pwr9 < %s | FileCheck %s --check-prefix=CHECK-PWR9
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-linux-gnu \
+; RUN:   -mcpu=pwr7 < %s | FileCheck %s
+; RUN: llc -verify-machineinstrs -mtriple=powerpc-unknown-aix \
+; RUN:   -mcpu=pwr7 < %s | FileCheck %s
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-aix \
+; RUN:   -mcpu=pwr8 < %s | FileCheck %s
+
+declare void @llvm.ppc.stfiw(i8*, double)
+define dso_local void @test_stfiw(i32* %cia, double %da) {
+; CHECK-LABEL: test_stfiw:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    stfiwx 1, 0, 3
+; CHECK-NEXT:    blr
+;
+; CHECK-NO-VSX-LABEL: test_stfiw:
+; CHECK-NO-VSX:       # %bb.0: # %entry
+; CHECK-NO-VSX-NEXT:    stfiwx 1, 0, 3
+; CHECK-NO-VSX-NEXT:    blr
+;
+; CHECK-PWR9-LABEL: test_stfiw:
+; CHECK-PWR9:       # %bb.0: # %entry
+; CHECK-PWR9-NEXT:    stxsiwx 1, 0, 3
+; CHECK-PWR9-NEXT:    blr
+entry:
+  %0 = bitcast i32* %cia to i8*
+  tail call void @llvm.ppc.stfiw(i8* %0, double %da)
+  ret void
+}
+
+define dso_local void @test_xl_stfiw(i32* %cia, double %da) {
+; CHECK-LABEL: test_xl_stfiw:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    stfiwx 1, 0, 3
+; CHECK-NEXT:    blr
+;
+; CHECK-NO-VSX-LABEL: test_xl_stfiw:
+; CHECK-NO-VSX:       # %bb.0: # %entry
+; CHECK-NO-VSX-NEXT:    stfiwx 1, 0, 3
+; CHECK-NO-VSX-NEXT:    blr
+;
+; CHECK-PWR9-LABEL: test_xl_stfiw:
+; CHECK-PWR9:       # %bb.0: # %entry
+; CHECK-PWR9-NEXT:    stxsiwx 1, 0, 3
+; CHECK-PWR9-NEXT:    blr
+entry:
+  %0 = bitcast i32* %cia to i8*
+  tail call void @llvm.ppc.stfiw(i8* %0, double %da)
+  ret void
+}
diff --git a/llvm/test/CodeGen/builtins-ppc-xlcompat-move-tofrom-regs.ll b/llvm/test/CodeGen/builtins-ppc-xlcompat-move-tofrom-regs.ll
new file mode 100644
--- /dev/null
+++ b/llvm/test/CodeGen/builtins-ppc-xlcompat-move-tofrom-regs.ll
@@ -0,0 +1,46 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu \
+; RUN:   -mcpu=pwr8 < %s | FileCheck %s
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-linux-gnu \
+; RUN:   -mcpu=pwr7 < %s | FileCheck %s
+; RUN: llc -verify-machineinstrs -mtriple=powerpc-unknown-aix \
+; RUN:   -mcpu=pwr7 < %s | FileCheck %s --check-prefix=CHECK-32BIT
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-aix \
+; RUN:   -mcpu=pwr7 < %s | FileCheck %s
+
+declare i32 @llvm.ppc.mftbu()
+declare i32 @llvm.ppc.mfmsr()
+
+define dso_local zeroext i32 @test_mftbu() {
+; CHECK-LABEL: test_mftbu:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    mftbu 3
+; CHECK-NEXT:    clrldi 3, 3, 32
+; CHECK-NEXT:    blr
+;
+; CHECK-32BIT-LABEL: test_mftbu:
+; CHECK-32BIT:       # %bb.0: # %entry
+; CHECK-32BIT-NEXT:    mftbu 3
+; CHECK-32BIT-NEXT:    blr
+entry:
+  %0 = tail call i32 @llvm.ppc.mftbu()
+  ret i32 %0
+}
+
+define dso_local i64 @test_mfmsr() {
+; CHECK-LABEL: test_mfmsr:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    mfmsr 3
+; CHECK-NEXT:    clrldi 3, 3, 32
+; CHECK-NEXT:    blr
+;
+; CHECK-32BIT-LABEL: test_mfmsr:
+; CHECK-32BIT:       # %bb.0: # %entry
+; CHECK-32BIT-NEXT:    mfmsr 4
+; CHECK-32BIT-NEXT:    li 3, 0
+; CHECK-32BIT-NEXT:    blr
+entry:
+  %0 = tail call i32 @llvm.ppc.mfmsr()
+  %conv = zext i32 %0 to i64
+  ret i64 %conv
+}
diff --git a/llvm/test/CodeGen/builtins-ppc-xlcompat-prefetch.ll b/llvm/test/CodeGen/builtins-ppc-xlcompat-prefetch.ll
new file mode 100644
--- /dev/null
+++ b/llvm/test/CodeGen/builtins-ppc-xlcompat-prefetch.ll
@@ -0,0 +1,71 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu \
+; RUN:   -mcpu=pwr8 < %s | FileCheck %s
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-linux-gnu \
+; RUN:   -mcpu=pwr7 < %s | FileCheck %s
+; RUN: llc -verify-machineinstrs -mtriple=powerpc-unknown-aix \
+; RUN:   -mcpu=pwr7 < %s | FileCheck %s --check-prefix=CHECK-AIX
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-aix \
+; RUN:   -mcpu=pwr7 < %s | FileCheck %s --check-prefix=CHECK-AIX64
+
+declare void @llvm.ppc.dcbtstt(i8*)
+declare void @llvm.ppc.dcbtt(i8*)
+
+@vpa = external local_unnamed_addr global i8*, align 8
+
+define dso_local void @test_dcbtstt() {
+; CHECK-LABEL: test_dcbtstt:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    addis 3, 2, .LC0@toc@ha
+; CHECK-NEXT:    ld 3, .LC0@toc@l(3)
+; CHECK-NEXT:    ld 3, 0(3)
+; CHECK-NEXT:    dcbtstt 0, 3
+; CHECK-NEXT:    blr
+;
+; CHECK-AIX-LABEL: test_dcbtstt:
+; CHECK-AIX:       # %bb.0: # %entry
+; CHECK-AIX-NEXT:    lwz 3, L..C0(2) # @vpa
+; CHECK-AIX-NEXT:    lwz 3, 0(3)
+; CHECK-AIX-NEXT:    dcbtstt 0, 3
+; CHECK-AIX-NEXT:    blr
+;
+; CHECK-AIX64-LABEL: test_dcbtstt:
+; CHECK-AIX64:       # %bb.0: # %entry
+; CHECK-AIX64-NEXT:    ld 3, L..C0(2) # @vpa
+; CHECK-AIX64-NEXT:    ld 3, 0(3)
+; CHECK-AIX64-NEXT:    dcbtstt 0, 3
+; CHECK-AIX64-NEXT:    blr
+entry:
+  %0 = load i8*, i8** @vpa, align 8
+  tail call void @llvm.ppc.dcbtstt(i8* %0)
+  ret void
+}
+
+
+define dso_local void @test_dcbtt() {
+; CHECK-LABEL: test_dcbtt:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    addis 3, 2, .LC0@toc@ha
+; CHECK-NEXT:    ld 3, .LC0@toc@l(3)
+; CHECK-NEXT:    ld 3, 0(3)
+; CHECK-NEXT:    dcbtt 0, 3
+; CHECK-NEXT:    blr
+;
+; CHECK-AIX-LABEL: test_dcbtt:
+; CHECK-AIX:       # %bb.0: # %entry
+; CHECK-AIX-NEXT:    lwz 3, L..C0(2) # @vpa
+; CHECK-AIX-NEXT:    lwz 3, 0(3)
+; CHECK-AIX-NEXT:    dcbtt 0, 3
+; CHECK-AIX-NEXT:    blr
+;
+; CHECK-AIX64-LABEL: test_dcbtt:
+; CHECK-AIX64:       # %bb.0: # %entry
+; CHECK-AIX64-NEXT:    ld 3, L..C0(2) # @vpa
+; CHECK-AIX64-NEXT:    ld 3, 0(3)
+; CHECK-AIX64-NEXT:    dcbtt 0, 3
+; CHECK-AIX64-NEXT:    blr
+entry:
+  %0 = load i8*, i8** @vpa, align 8
+  tail call void @llvm.ppc.dcbtt(i8* %0)
+  ret void
+}