diff --git a/llvm/lib/Target/PowerPC/PPCISelLowering.cpp b/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
index b29674d992bb..2808b41671d1 100644
--- a/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
+++ b/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
@@ -8312,6 +8312,8 @@ SDValue PPCTargetLowering::PerformDAGCombine(SDNode *N,
     unsigned ABIAlignment = getDataLayout()->getABITypeAlignment(Ty);
     if (ISD::isNON_EXTLoad(N) && VT.isVector() &&
         TM.getSubtarget<PPCSubtarget>().hasAltivec() &&
+        // P8 and later hardware should just use LOAD.
+        !TM.getSubtarget<PPCSubtarget>().hasP8Vector() &&
         (VT == MVT::v16i8 || VT == MVT::v8i16 ||
          VT == MVT::v4i32 || VT == MVT::v4f32) &&
         LD->getAlignment() < ABIAlignment) {
@@ -9204,7 +9206,8 @@ bool PPCTargetLowering::allowsMisalignedMemoryAccesses(EVT VT,
 
   if (VT.getSimpleVT().isVector()) {
     if (Subtarget.hasVSX()) {
-      if (VT != MVT::v2f64 && VT != MVT::v2i64)
+      if (VT != MVT::v2f64 && VT != MVT::v2i64 &&
+          VT != MVT::v4f32 && VT != MVT::v4i32)
         return false;
     } else {
       return false;
diff --git a/llvm/lib/Target/PowerPC/PPCInstrVSX.td b/llvm/lib/Target/PowerPC/PPCInstrVSX.td
index f30e64f95d17..4b2f5a321f80 100644
--- a/llvm/lib/Target/PowerPC/PPCInstrVSX.td
+++ b/llvm/lib/Target/PowerPC/PPCInstrVSX.td
@@ -63,7 +63,8 @@ let Uses = [RM] in {
 
     def LXVW4X : XX1Form<31, 780,
                          (outs vsrc:$XT), (ins memrr:$src),
-                         "lxvw4x $XT, $src", IIC_LdStLFD, []>;
+                         "lxvw4x $XT, $src", IIC_LdStLFD,
+                         [(set v4i32:$XT, (load xoaddr:$src))]>;
   }
 
   // Store indexed instructions
@@ -80,7 +81,8 @@ let Uses = [RM] in {
 
     def STXVW4X : XX1Form<31, 908,
                          (outs), (ins vsrc:$XT, memrr:$dst),
-                         "stxvw4x $XT, $dst", IIC_LdStSTFD, []>;
+                         "stxvw4x $XT, $dst", IIC_LdStSTFD,
+                         [(store v4i32:$XT, xoaddr:$dst)]>;
   }
 
   // Add/Mul Instructions
@@ -811,6 +813,13 @@ def : Pat<(sext_inreg v2i64:$C, v2i32),
 def : Pat<(v2f64 (sint_to_fp (sext_inreg v2i64:$C, v2i32))),
           (XVCVSXWDP (XXSLDWI $C, $C, 1))>;
 
+// Loads.
+def : Pat<(v4i32 (load xoaddr:$src)), (LXVW4X xoaddr:$src)>;
+
+// Stores.
+def : Pat<(store v4i32:$rS, xoaddr:$dst),
+          (STXVW4X $rS, xoaddr:$dst)>;
+
 } // AddedComplexity
 } // HasVSX
 
diff --git a/llvm/test/CodeGen/PowerPC/unal4-std.ll b/llvm/test/CodeGen/PowerPC/unal4-std.ll
index 9f29e31cb902..e91109911161 100644
--- a/llvm/test/CodeGen/PowerPC/unal4-std.ll
+++ b/llvm/test/CodeGen/PowerPC/unal4-std.ll
@@ -1,4 +1,5 @@
-; RUN: llc < %s -mcpu=pwr7 | FileCheck %s
+; RUN: llc < %s -mcpu=pwr7 -mattr=-vsx| FileCheck %s
+; RUN: llc < %s -mcpu=pwr7 -mattr=+vsx | FileCheck -check-prefix=CHECK-VSX %s
 target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-f128:128:128-v128:128:128-n32:64"
 target triple = "powerpc64-unknown-linux-gnu"
 
@@ -22,6 +23,9 @@ if.end210:                                        ; preds = %entry
 ; a multiple of 4).
 ; CHECK: @copy_to_conceal
 ; CHECK: stdx {{[0-9]+}}, 0,
+
+; CHECK-VSX: @copy_to_conceal
+; CHECK-VSX: stxvw4x {{[0-9]+}}, 0,
 }
 
 attributes #0 = { nounwind "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-frame-pointer-elim-non-leaf" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "unsafe-fp-math"="false" "use-soft-float"="false" }
diff --git a/llvm/test/CodeGen/PowerPC/unaligned.ll b/llvm/test/CodeGen/PowerPC/unaligned.ll
index 0c59516f1186..64c03cdda35e 100644
--- a/llvm/test/CodeGen/PowerPC/unaligned.ll
+++ b/llvm/test/CodeGen/PowerPC/unaligned.ll
@@ -92,10 +92,14 @@ entry:
 ; CHECK-DAG: stdx
 ; CHECK: stdx
 
+; For VSX on P7, unaligned loads and stores are preferable to aligned
+; stack slots, but lvsl/vperm is better still.  (On P8 lxvw4x is preferable.)
+; Using unaligned stxvw4x is preferable on both machines.
 ; CHECK-VSX: @foo6
-; CHECK-VSX-DAG: ld
-; CHECK-VSX-DAG: ld
-; CHECK-VSX-DAG: stdx
-; CHECK-VSX: stdx
+; CHECK-VSX-DAG: lvsl
+; CHECK-VSX-DAG: lvx
+; CHECK-VSX-DAG: lvx
+; CHECK-VSX: vperm
+; CHECK-VSX: stxvw4x
 }
 
diff --git a/llvm/test/CodeGen/PowerPC/vec-abi-align.ll b/llvm/test/CodeGen/PowerPC/vec-abi-align.ll
index 3239cf6c06ab..5075ff2b8c07 100644
--- a/llvm/test/CodeGen/PowerPC/vec-abi-align.ll
+++ b/llvm/test/CodeGen/PowerPC/vec-abi-align.ll
@@ -1,4 +1,5 @@
-; RUN: llc -mtriple=powerpc64-unknown-linux-gnu -mcpu=pwr7 < %s | FileCheck %s
+; RUN: llc -mtriple=powerpc64-unknown-linux-gnu -mcpu=pwr7 -mattr=-vsx < %s | FileCheck %s
+; RUN: llc -mtriple=powerpc64-unknown-linux-gnu -mcpu=pwr7 -mattr=+vsx < %s | FileCheck -check-prefix=CHECK-VSX %s
 target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-f128:128:128-v128:128:128-n32:64"
 target triple = "powerpc64-unknown-linux-gnu"
 
@@ -16,6 +17,10 @@ entry:
 ; CHECK-LABEL: @test1
 ; CHECK: stvx 2,
 ; CHECK: blr
+
+; CHECK-VSX-LABEL: @test1
+; CHECK-VSX: stxvw4x 34,
+; CHECK-VSX: blr
 }
 
 ; Function Attrs: nounwind
@@ -35,6 +40,13 @@ entry:
 ; CHECK: addi [[REGB:[0-9]+]], 1, 112
 ; CHECK: lvx 2, [[REGB]], [[REG16]]
 ; CHECK: blr
+
+; CHECK-VSX-LABEL: @test2
+; CHECK-VSX: ld {{[0-9]+}}, 112(1)
+; CHECK-VSX: li [[REG16:[0-9]+]], 16
+; CHECK-VSX: addi [[REGB:[0-9]+]], 1, 112
+; CHECK-VSX: lxvw4x {{[0-9]+}}, [[REGB]], [[REG16]]
+; CHECK-VSX: blr
 }
 
 ; Function Attrs: nounwind
@@ -54,6 +66,13 @@ entry:
 ; CHECK: addi [[REGB:[0-9]+]], 1, 128
 ; CHECK: lvx 2, [[REGB]], [[REG16]]
 ; CHECK: blr
+
+; CHECK-VSX-LABEL: @test3
+; CHECK-VSX: ld {{[0-9]+}}, 128(1)
+; CHECK-VSX: li [[REG16:[0-9]+]], 16
+; CHECK-VSX: addi [[REGB:[0-9]+]], 1, 128
+; CHECK-VSX: lxvw4x {{[0-9]+}}, [[REGB]], [[REG16]]
+; CHECK-VSX: blr
 }
 
 attributes #0 = { nounwind }
diff --git a/llvm/test/CodeGen/PowerPC/vec_misaligned.ll b/llvm/test/CodeGen/PowerPC/vec_misaligned.ll
index 304a84d49a9d..73a4a4d395da 100644
--- a/llvm/test/CodeGen/PowerPC/vec_misaligned.ll
+++ b/llvm/test/CodeGen/PowerPC/vec_misaligned.ll
@@ -1,5 +1,5 @@
 ; RUN: llc < %s -march=ppc32 -mcpu=g5 | FileCheck %s
-; RUN: llc < %s -mtriple=powerpc64-unknown-linux-gnu -mattr=+altivec | FileCheck %s
+; RUN: llc < %s -mtriple=powerpc64-unknown-linux-gnu -mattr=+altivec -mattr=-vsx -mattr=-power8-vector | FileCheck %s
 ; RUN: llc < %s -mtriple=powerpc64le-unknown-linux-gnu -mattr=+altivec | FileCheck %s -check-prefix=CHECK-LE
 
 target datalayout = "E-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f128:64:128"
diff --git a/llvm/test/CodeGen/PowerPC/vrspill.ll b/llvm/test/CodeGen/PowerPC/vrspill.ll
index c3d1bf8f1ead..b990442aed87 100644
--- a/llvm/test/CodeGen/PowerPC/vrspill.ll
+++ b/llvm/test/CodeGen/PowerPC/vrspill.ll
@@ -1,5 +1,7 @@
-; RUN: llc -O0 -mtriple=powerpc-unknown-linux-gnu -mattr=+altivec -verify-machineinstrs < %s | FileCheck %s
-; RUN: llc -O0 -mtriple=powerpc64-unknown-linux-gnu -mattr=+altivec -verify-machineinstrs -fast-isel=false < %s | FileCheck %s
+; RUN: llc -O0 -mtriple=powerpc-unknown-linux-gnu -mattr=+altivec -mattr=-vsx -verify-machineinstrs < %s | FileCheck %s
+; RUN: llc -O0 -mtriple=powerpc64-unknown-linux-gnu -mattr=+altivec -mattr=-vsx -verify-machineinstrs -fast-isel=false < %s | FileCheck %s
+; RUN: llc -O0 -mtriple=powerpc-unknown-linux-gnu -mattr=+altivec -mattr=+vsx -verify-machineinstrs < %s | FileCheck -check-prefix=CHECK-VSX %s
+; RUN: llc -O0 -mtriple=powerpc64-unknown-linux-gnu -mattr=+altivec -mattr=+vsx -verify-machineinstrs -fast-isel=false < %s | FileCheck -check-prefix=CHECK-VSX %s
 
 ; This verifies that we generate correct spill/reload code for vector regs.
 
@@ -15,4 +17,9 @@ entry:
 
 ; CHECK: stvx 2,
 
+; We would prefer to test for "stxvw4x 34," but current -O0 code
+; needlessly generates "vor 3,2,2 / stxvw4x 35,0,3", so we'll settle for
+; the opcode.
+; CHECK-VSX: stxvw4x
+
 declare void @foo(i32*)
diff --git a/llvm/test/CodeGen/PowerPC/vsx-p8.ll b/llvm/test/CodeGen/PowerPC/vsx-p8.ll
new file mode 100644
index 000000000000..81406b6f0798
--- /dev/null
+++ b/llvm/test/CodeGen/PowerPC/vsx-p8.ll
@@ -0,0 +1,42 @@
+; RUN: llc -mcpu=pwr8 -mattr=+power8-vector < %s | FileCheck %s
+target datalayout = "E-m:e-i64:64-n32:64"
+target triple = "powerpc64-unknown-linux-gnu"
+
+; Unaligned loads/stores on P8 and later should use VSX where possible.
+
+define <2 x double> @test28u(<2 x double>* %a) {
+  %v = load <2 x double>* %a, align 8
+  ret <2 x double> %v
+
+; CHECK-LABEL: @test28u
+; CHECK: lxvd2x 34, 0, 3
+; CHECK: blr
+}
+
+define void @test29u(<2 x double>* %a, <2 x double> %b) {
+  store <2 x double> %b, <2 x double>* %a, align 8
+  ret void
+
+; CHECK-LABEL: @test29u
+; CHECK: stxvd2x 34, 0, 3
+; CHECK: blr
+}
+
+define <4 x float> @test32u(<4 x float>* %a) {
+  %v = load <4 x float>* %a, align 8
+  ret <4 x float> %v
+
+; CHECK-LABEL: @test32u
+; CHECK: lxvw4x 34, 0, 3
+; CHECK: blr
+}
+
+define void @test33u(<4 x float>* %a, <4 x float> %b) {
+  store <4 x float> %b, <4 x float>* %a, align 8
+  ret void
+
+; CHECK-LABEL: @test33u
+; CHECK: stxvw4x 34, 0, 3
+; CHECK: blr
+}
+
diff --git a/llvm/test/CodeGen/PowerPC/vsx.ll b/llvm/test/CodeGen/PowerPC/vsx.ll
index 2f226e1f614c..65343f4a9ba6 100644
--- a/llvm/test/CodeGen/PowerPC/vsx.ll
+++ b/llvm/test/CodeGen/PowerPC/vsx.ll
@@ -356,6 +356,63 @@ define void @test31(<2 x i64>* %a, <2 x i64> %b) {
 ; CHECK: blr
 }
 
+define <4 x float> @test32(<4 x float>* %a) {
+  %v = load <4 x float>* %a, align 16
+  ret <4 x float> %v
+
+; CHECK-LABEL: @test32
+; CHECK: lxvw4x 34, 0, 3
+; CHECK: blr
+}
+
+define void @test33(<4 x float>* %a, <4 x float> %b) {
+  store <4 x float> %b, <4 x float>* %a, align 16
+  ret void
+
+; CHECK-LABEL: @test33
+; CHECK: stxvw4x 34, 0, 3
+; CHECK: blr
+}
+
+define <4 x float> @test32u(<4 x float>* %a) {
+  %v = load <4 x float>* %a, align 8
+  ret <4 x float> %v
+
+; CHECK-LABEL: @test32u
+; CHECK-DAG: lvsl
+; CHECK-DAG: lvx
+; CHECK-DAG: lvx
+; CHECK: vperm 2,
+; CHECK: blr
+}
+
+define void @test33u(<4 x float>* %a, <4 x float> %b) {
+  store <4 x float> %b, <4 x float>* %a, align 8
+  ret void
+
+; CHECK-LABEL: @test33u
+; CHECK: stxvw4x 34, 0, 3
+; CHECK: blr
+}
+
+define <4 x i32> @test34(<4 x i32>* %a) {
+  %v = load <4 x i32>* %a, align 16
+  ret <4 x i32> %v
+
+; CHECK-LABEL: @test34
+; CHECK: lxvw4x 34, 0, 3
+; CHECK: blr
+}
+
+define void @test35(<4 x i32>* %a, <4 x i32> %b) {
+  store <4 x i32> %b, <4 x i32>* %a, align 16
+  ret void
+
+; CHECK-LABEL: @test35
+; CHECK: stxvw4x 34, 0, 3
+; CHECK: blr
+}
+
 define <2 x double> @test40(<2 x i64> %a) {
   %v = uitofp <2 x i64> %a to <2 x double>
   ret <2 x double> %v