From 0c1000cbd6d25d749c78f9c27fa985a2608ff217 Mon Sep 17 00:00:00 2001
From: Amy Kwan <amy.kwan1@ibm.com>
Date: Mon, 9 May 2022 10:51:08 -0500
Subject: [PATCH] [NFC][PowerPC] Add 32-bit AIX RUN lines to test cases.

This patch adds 32-bit AIX RUN lines to several test cases, along with the
addition of one new test case, to prepare for future codegen changes involving
the PPCISD::SCALAR_TO_VECTOR_PERMUTED node on 32-bit mode.
---
 .../PowerPC/aix_scalar_vector_permuted.ll     |  86 ++++
 .../PowerPC/canonical-merge-shuffles.ll       | 259 ++++++++++
 .../CodeGen/PowerPC/float-vector-gather.ll    |  21 +
 .../CodeGen/PowerPC/load-v4i8-improved.ll     |  18 +-
 llvm/test/CodeGen/PowerPC/pre-inc-disable.ll  |  19 +
 .../CodeGen/PowerPC/reduce_scalarization.ll   | 166 +++++++
 .../CodeGen/PowerPC/scalar_vector_test_4.ll   | 292 +++++++++++
 llvm/test/CodeGen/PowerPC/vec_insert_elt.ll   | 453 ++++++++++++++++++
 8 files changed, 1313 insertions(+), 1 deletion(-)
 create mode 100644 llvm/test/CodeGen/PowerPC/aix_scalar_vector_permuted.ll

diff --git a/llvm/test/CodeGen/PowerPC/aix_scalar_vector_permuted.ll b/llvm/test/CodeGen/PowerPC/aix_scalar_vector_permuted.ll
new file mode 100644
index 000000000000..71d70167b12c
--- /dev/null
+++ b/llvm/test/CodeGen/PowerPC/aix_scalar_vector_permuted.ll
@@ -0,0 +1,86 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mcpu=pwr8 -verify-machineinstrs -ppc-vsr-nums-as-vr \
+; RUN:   -ppc-asm-full-reg-names -mtriple=powerpc64-ibm-aix-xcoff < %s | \
+; RUN: FileCheck %s --check-prefix=AIX-P8-64
+; RUN: llc -mcpu=pwr8 -verify-machineinstrs -ppc-vsr-nums-as-vr \
+; RUN:   -ppc-asm-full-reg-names -mtriple=powerpc-ibm-aix-xcoff < %s | \
+; RUN: FileCheck %s --check-prefix=AIX-P8-32
+; RUN: llc -mcpu=pwr9 -verify-machineinstrs -ppc-vsr-nums-as-vr \
+; RUN:   -ppc-asm-full-reg-names -mtriple=powerpc64-ibm-aix-xcoff < %s | \
+; RUN: FileCheck %s --check-prefix=AIX-P9-64
+; RUN: llc -mcpu=pwr9 -verify-machineinstrs -ppc-vsr-nums-as-vr \
+; RUN:   -ppc-asm-full-reg-names -mtriple=powerpc-ibm-aix-xcoff < %s | \
+; RUN: FileCheck %s --check-prefix=AIX-P9-32
+
+%d8 = type <8 x double>
+%f1 = type <1 x float>
+%f2 = type <2 x float>
+%f4 = type <4 x float>
+%f8 = type <8 x float>
+%i4 = type <4 x i32>
+
+define void @test_f2(%f2* %P, %f2* %Q, %f2* %S) {
+; AIX-P8-64-LABEL: test_f2:
+; AIX-P8-64:       # %bb.0:
+; AIX-P8-64-NEXT:    lfdx f0, 0, r3
+; AIX-P8-64-NEXT:    lfdx f1, 0, r4
+; AIX-P8-64-NEXT:    xvaddsp vs0, vs0, vs1
+; AIX-P8-64-NEXT:    stfdx f0, 0, r5
+; AIX-P8-64-NEXT:    blr
+;
+; AIX-P8-32-LABEL: test_f2:
+; AIX-P8-32:       # %bb.0:
+; AIX-P8-32-NEXT:    lfs f0, 4(r3)
+; AIX-P8-32-NEXT:    lfs f1, 0(r3)
+; AIX-P8-32-NEXT:    lwz r6, L..C0(r2) # %const.0
+; AIX-P8-32-NEXT:    lfs f2, 4(r4)
+; AIX-P8-32-NEXT:    xscvdpspn v2, f0
+; AIX-P8-32-NEXT:    lfs f0, 0(r4)
+; AIX-P8-32-NEXT:    lxvw4x v0, 0, r6
+; AIX-P8-32-NEXT:    xscvdpspn v3, f1
+; AIX-P8-32-NEXT:    xscvdpspn v4, f2
+; AIX-P8-32-NEXT:    xscvdpspn v5, f0
+; AIX-P8-32-NEXT:    vperm v2, v3, v2, v0
+; AIX-P8-32-NEXT:    vperm v3, v5, v4, v0
+; AIX-P8-32-NEXT:    xvaddsp vs0, v2, v3
+; AIX-P8-32-NEXT:    xxsldwi vs1, vs0, vs0, 1
+; AIX-P8-32-NEXT:    xscvspdpn f0, vs0
+; AIX-P8-32-NEXT:    xscvspdpn f1, vs1
+; AIX-P8-32-NEXT:    stfs f0, 0(r5)
+; AIX-P8-32-NEXT:    stfs f1, 4(r5)
+; AIX-P8-32-NEXT:    blr
+;
+; AIX-P9-64-LABEL: test_f2:
+; AIX-P9-64:       # %bb.0:
+; AIX-P9-64-NEXT:    lfd f0, 0(r3)
+; AIX-P9-64-NEXT:    lfd f1, 0(r4)
+; AIX-P9-64-NEXT:    xvaddsp vs0, vs0, vs1
+; AIX-P9-64-NEXT:    stfd f0, 0(r5)
+; AIX-P9-64-NEXT:    blr
+;
+; AIX-P9-32-LABEL: test_f2:
+; AIX-P9-32:       # %bb.0:
+; AIX-P9-32-NEXT:    lfs f0, 0(r3)
+; AIX-P9-32-NEXT:    lwz r3, 4(r3)
+; AIX-P9-32-NEXT:    mtfprwz f1, r3
+; AIX-P9-32-NEXT:    lwz r3, 4(r4)
+; AIX-P9-32-NEXT:    xscvdpspn vs0, f0
+; AIX-P9-32-NEXT:    mtfprwz f2, r3
+; AIX-P9-32-NEXT:    xxinsertw vs0, vs1, 4
+; AIX-P9-32-NEXT:    lfs f1, 0(r4)
+; AIX-P9-32-NEXT:    xscvdpspn vs1, f1
+; AIX-P9-32-NEXT:    xxinsertw vs1, vs2, 4
+; AIX-P9-32-NEXT:    xvaddsp vs0, vs0, vs1
+; AIX-P9-32-NEXT:    xscvspdpn f1, vs0
+; AIX-P9-32-NEXT:    xxsldwi vs0, vs0, vs0, 1
+; AIX-P9-32-NEXT:    xscvspdpn f0, vs0
+; AIX-P9-32-NEXT:    stfs f1, 0(r5)
+; AIX-P9-32-NEXT:    stfs f0, 4(r5)
+; AIX-P9-32-NEXT:    blr
+  %p = load %f2, %f2* %P
+  %q = load %f2, %f2* %Q
+  %R = fadd %f2 %p, %q
+  store %f2 %R, %f2* %S
+  ret void
+}
+
diff --git a/llvm/test/CodeGen/PowerPC/canonical-merge-shuffles.ll b/llvm/test/CodeGen/PowerPC/canonical-merge-shuffles.ll
index 12c2505f2b2e..1d424446db7c 100644
--- a/llvm/test/CodeGen/PowerPC/canonical-merge-shuffles.ll
+++ b/llvm/test/CodeGen/PowerPC/canonical-merge-shuffles.ll
@@ -14,6 +14,12 @@
 ; RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu \
 ; RUN:     -mcpu=pwr7 -ppc-asm-full-reg-names -ppc-vsr-nums-as-vr < %s | \
 ; RUN: FileCheck %s --check-prefix=CHECK-P7
+; RUN: llc -mcpu=pwr8 -verify-machineinstrs -ppc-vsr-nums-as-vr \
+; RUN:   -ppc-asm-full-reg-names -mtriple=powerpc64-ibm-aix-xcoff < %s | \
+; RUN: FileCheck %s --check-prefixes=P8-AIX,P8-AIX-64
+; RUN: llc -mcpu=pwr8 -verify-machineinstrs -ppc-vsr-nums-as-vr \
+; RUN:   -ppc-asm-full-reg-names -mtriple=powerpc-ibm-aix-xcoff < %s | \
+; RUN: FileCheck %s --check-prefixes=P8-AIX,P8-AIX-32
 
 define dso_local <16 x i8> @testmrghb(<16 x i8> %a, <16 x i8> %b) local_unnamed_addr #0 {
 ; CHECK-P8-LABEL: testmrghb:
@@ -40,6 +46,11 @@ define dso_local <16 x i8> @testmrghb(<16 x i8> %a, <16 x i8> %b) local_unnamed_
 ; CHECK-P7:       # %bb.0: # %entry
 ; CHECK-P7-NEXT:    vmrghb v2, v3, v2
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-LABEL: testmrghb:
+; P8-AIX:       # %bb.0: # %entry
+; P8-AIX-NEXT:    vmrglb v2, v2, v3
+; P8-AIX-NEXT:    blr
 entry:
   %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
   ret <16 x i8> %shuffle
@@ -75,6 +86,11 @@ define dso_local <16 x i8> @testmrghb2(<16 x i8> %a, <16 x i8> %b) local_unnamed
 ; CHECK-P7-NEXT:    lvx v4, 0, r3
 ; CHECK-P7-NEXT:    vperm v2, v3, v2, v4
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-LABEL: testmrghb2:
+; P8-AIX:       # %bb.0: # %entry
+; P8-AIX-NEXT:    vmrglb v2, v3, v2
+; P8-AIX-NEXT:    blr
 entry:
   %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 24, i32 8, i32 25, i32 9, i32 26, i32 10, i32 27, i32 11, i32 28, i32 12, i32 29, i32 13, i32 30, i32 14, i32 31, i32 15>
   ret <16 x i8> %shuffle
@@ -104,6 +120,11 @@ define dso_local <16 x i8> @testmrghh(<16 x i8> %a, <16 x i8> %b) local_unnamed_
 ; CHECK-P7:       # %bb.0: # %entry
 ; CHECK-P7-NEXT:    vmrghh v2, v3, v2
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-LABEL: testmrghh:
+; P8-AIX:       # %bb.0: # %entry
+; P8-AIX-NEXT:    vmrglh v2, v2, v3
+; P8-AIX-NEXT:    blr
 entry:
   %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 8, i32 9, i32 24, i32 25, i32 10, i32 11, i32 26, i32 27, i32 12, i32 13, i32 28, i32 29, i32 14, i32 15, i32 30, i32 31>
   ret <16 x i8> %shuffle
@@ -139,6 +160,11 @@ define dso_local <16 x i8> @testmrghh2(<16 x i8> %a, <16 x i8> %b) local_unnamed
 ; CHECK-P7-NEXT:    lvx v4, 0, r3
 ; CHECK-P7-NEXT:    vperm v2, v3, v2, v4
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-LABEL: testmrghh2:
+; P8-AIX:       # %bb.0: # %entry
+; P8-AIX-NEXT:    vmrglh v2, v3, v2
+; P8-AIX-NEXT:    blr
 entry:
   %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 24, i32 25, i32 8, i32 9, i32 26, i32 27, i32 10, i32 11, i32 28, i32 29, i32 12, i32 13, i32 30, i32 31, i32 14, i32 15>
   ret <16 x i8> %shuffle
@@ -168,6 +194,11 @@ define dso_local <16 x i8> @testmrglb(<16 x i8> %a, <16 x i8> %b) local_unnamed_
 ; CHECK-P7:       # %bb.0: # %entry
 ; CHECK-P7-NEXT:    vmrglb v2, v3, v2
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-LABEL: testmrglb:
+; P8-AIX:       # %bb.0: # %entry
+; P8-AIX-NEXT:    vmrghb v2, v2, v3
+; P8-AIX-NEXT:    blr
 entry:
   %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
   ret <16 x i8> %shuffle
@@ -203,6 +234,11 @@ define dso_local <16 x i8> @testmrglb2(<16 x i8> %a, <16 x i8> %b) local_unnamed
 ; CHECK-P7-NEXT:    lvx v4, 0, r3
 ; CHECK-P7-NEXT:    vperm v2, v3, v2, v4
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-LABEL: testmrglb2:
+; P8-AIX:       # %bb.0: # %entry
+; P8-AIX-NEXT:    vmrghb v2, v3, v2
+; P8-AIX-NEXT:    blr
 entry:
   %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 16, i32 0, i32 17, i32 1, i32 18, i32 2, i32 19, i32 3, i32 20, i32 4, i32 21, i32 5, i32 22, i32 6, i32 23, i32 7>
   ret <16 x i8> %shuffle
@@ -232,6 +268,11 @@ define dso_local <16 x i8> @testmrglh(<16 x i8> %a, <16 x i8> %b) local_unnamed_
 ; CHECK-P7:       # %bb.0: # %entry
 ; CHECK-P7-NEXT:    vmrglh v2, v3, v2
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-LABEL: testmrglh:
+; P8-AIX:       # %bb.0: # %entry
+; P8-AIX-NEXT:    vmrghh v2, v2, v3
+; P8-AIX-NEXT:    blr
 entry:
   %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 1, i32 16, i32 17, i32 2, i32 3, i32 18, i32 19, i32 4, i32 5, i32 20, i32 21, i32 6, i32 7, i32 22, i32 23>
   ret <16 x i8> %shuffle
@@ -267,6 +308,11 @@ define dso_local <16 x i8> @testmrglh2(<16 x i8> %a, <16 x i8> %b) local_unnamed
 ; CHECK-P7-NEXT:    lvx v4, 0, r3
 ; CHECK-P7-NEXT:    vperm v2, v3, v2, v4
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-LABEL: testmrglh2:
+; P8-AIX:       # %bb.0: # %entry
+; P8-AIX-NEXT:    vmrghh v2, v3, v2
+; P8-AIX-NEXT:    blr
 entry:
   %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 16, i32 17, i32 0, i32 1, i32 18, i32 19, i32 2, i32 3, i32 20, i32 21, i32 4, i32 5, i32 22, i32 23, i32 6, i32 7>
   ret <16 x i8> %shuffle
@@ -296,6 +342,11 @@ define dso_local <16 x i8> @testmrghw(<16 x i8> %a, <16 x i8> %b) local_unnamed_
 ; CHECK-P7:       # %bb.0: # %entry
 ; CHECK-P7-NEXT:    xxmrghw v2, v3, v2
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-LABEL: testmrghw:
+; P8-AIX:       # %bb.0: # %entry
+; P8-AIX-NEXT:    xxmrglw v2, v2, v3
+; P8-AIX-NEXT:    blr
 entry:
   %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 24, i32 25, i32 26, i32 27, i32 12, i32 13, i32 14, i32 15, i32 28, i32 29, i32 30, i32 31>
   ret <16 x i8> %shuffle
@@ -331,6 +382,11 @@ define dso_local <16 x i8> @testmrghw2(<16 x i8> %a, <16 x i8> %b) local_unnamed
 ; CHECK-P7-NEXT:    lvx v4, 0, r3
 ; CHECK-P7-NEXT:    vperm v2, v3, v2, v4
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-LABEL: testmrghw2:
+; P8-AIX:       # %bb.0: # %entry
+; P8-AIX-NEXT:    xxmrglw v2, v3, v2
+; P8-AIX-NEXT:    blr
 entry:
   %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 24, i32 25, i32 26, i32 27, i32 8, i32 9, i32 10, i32 11, i32 28, i32 29, i32 30, i32 31, i32 12, i32 13, i32 14, i32 15>
   ret <16 x i8> %shuffle
@@ -360,6 +416,11 @@ define dso_local <16 x i8> @testmrglw(<16 x i8> %a, <16 x i8> %b) local_unnamed_
 ; CHECK-P7:       # %bb.0: # %entry
 ; CHECK-P7-NEXT:    xxmrglw v2, v3, v2
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-LABEL: testmrglw:
+; P8-AIX:       # %bb.0: # %entry
+; P8-AIX-NEXT:    xxmrghw v2, v2, v3
+; P8-AIX-NEXT:    blr
 entry:
   %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 4, i32 5, i32 6, i32 7, i32 20, i32 21, i32 22, i32 23>
   ret <16 x i8> %shuffle
@@ -395,6 +456,11 @@ define dso_local <16 x i8> @testmrglw2(<16 x i8> %a, <16 x i8> %b) local_unnamed
 ; CHECK-P7-NEXT:    lvx v4, 0, r3
 ; CHECK-P7-NEXT:    vperm v2, v3, v2, v4
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-LABEL: testmrglw2:
+; P8-AIX:       # %bb.0: # %entry
+; P8-AIX-NEXT:    xxmrghw v2, v3, v2
+; P8-AIX-NEXT:    blr
 entry:
   %shuffle = shufflevector <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 0, i32 1, i32 2, i32 3, i32 20, i32 21, i32 22, i32 23, i32 4, i32 5, i32 6, i32 7>
   ret <16 x i8> %shuffle
@@ -451,6 +517,32 @@ define dso_local <8 x i16> @testmrglb3(<8 x i8>* nocapture readonly %a) local_un
 ; CHECK-P7-NEXT:    xxswapd v2, vs0
 ; CHECK-P7-NEXT:    vperm v2, v2, v4, v3
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-64-LABEL: testmrglb3:
+; P8-AIX-64:       # %bb.0: # %entry
+; P8-AIX-64-NEXT:    ld r4, L..C0(r2) # %const.0
+; P8-AIX-64-NEXT:    lxsdx v2, 0, r3
+; P8-AIX-64-NEXT:    xxlxor v4, v4, v4
+; P8-AIX-64-NEXT:    lxvw4x v3, 0, r4
+; P8-AIX-64-NEXT:    vperm v2, v4, v2, v3
+; P8-AIX-64-NEXT:    blr
+;
+; P8-AIX-32-LABEL: testmrglb3:
+; P8-AIX-32:       # %bb.0: # %entry
+; P8-AIX-32-NEXT:    lwz r5, 4(r3)
+; P8-AIX-32-NEXT:    lwz r4, L..C0(r2) # %const.0
+; P8-AIX-32-NEXT:    stw r5, -32(r1)
+; P8-AIX-32-NEXT:    lwz r3, 0(r3)
+; P8-AIX-32-NEXT:    lxvw4x v2, 0, r4
+; P8-AIX-32-NEXT:    addi r4, r1, -16
+; P8-AIX-32-NEXT:    stw r3, -16(r1)
+; P8-AIX-32-NEXT:    addi r3, r1, -32
+; P8-AIX-32-NEXT:    lxvw4x v3, 0, r3
+; P8-AIX-32-NEXT:    lxvw4x v4, 0, r4
+; P8-AIX-32-NEXT:    vperm v2, v4, v3, v2
+; P8-AIX-32-NEXT:    xxlxor v3, v3, v3
+; P8-AIX-32-NEXT:    vmrghb v2, v3, v2
+; P8-AIX-32-NEXT:    blr
 entry:
   %0 = load <8 x i8>, <8 x i8>* %a, align 8
   %1 = zext <8 x i8> %0 to <8 x i16>
@@ -528,6 +620,34 @@ define dso_local void @no_crash_elt0_from_RHS(<2 x double>* noalias nocapture de
 ; CHECK-P7-NEXT:    xxmrghd vs0, vs1, vs0
 ; CHECK-P7-NEXT:    xxswapd vs0, vs0
 ; CHECK-P7-NEXT:    stxvd2x vs0, 0, r30
+;
+; P8-AIX-64-LABEL: no_crash_elt0_from_RHS:
+; P8-AIX-64:       # %bb.0: # %test_entry
+; P8-AIX-64-NEXT:    mflr r0
+; P8-AIX-64-NEXT:    std r0, 16(r1)
+; P8-AIX-64-NEXT:    stdu r1, -128(r1)
+; P8-AIX-64-NEXT:    std r31, 120(r1) # 8-byte Folded Spill
+; P8-AIX-64-NEXT:    mr r31, r3
+; P8-AIX-64-NEXT:    bl .dummy[PR]
+; P8-AIX-64-NEXT:    nop
+; P8-AIX-64-NEXT:    xxlxor f0, f0, f0
+; P8-AIX-64-NEXT:    # kill: def $f1 killed $f1 def $vsl1
+; P8-AIX-64-NEXT:    xxmrghd vs0, vs0, vs1
+; P8-AIX-64-NEXT:    stxvd2x vs0, 0, r31
+;
+; P8-AIX-32-LABEL: no_crash_elt0_from_RHS:
+; P8-AIX-32:       # %bb.0: # %test_entry
+; P8-AIX-32-NEXT:    mflr r0
+; P8-AIX-32-NEXT:    stw r0, 8(r1)
+; P8-AIX-32-NEXT:    stwu r1, -64(r1)
+; P8-AIX-32-NEXT:    stw r31, 60(r1) # 4-byte Folded Spill
+; P8-AIX-32-NEXT:    mr r31, r3
+; P8-AIX-32-NEXT:    bl .dummy[PR]
+; P8-AIX-32-NEXT:    nop
+; P8-AIX-32-NEXT:    xxlxor f0, f0, f0
+; P8-AIX-32-NEXT:    # kill: def $f1 killed $f1 def $vsl1
+; P8-AIX-32-NEXT:    xxmrghd vs0, vs0, vs1
+; P8-AIX-32-NEXT:    stxvd2x vs0, 0, r31
 test_entry:
   %_div_result = tail call double @dummy()
   %oldret = insertvalue { double, double } undef, double %_div_result, 0
@@ -577,6 +697,22 @@ define dso_local <16 x i8> @no_crash_bitcast(i32 %a) {
 ; CHECK-P7-NEXT:    lvx v2, 0, r4
 ; CHECK-P7-NEXT:    vperm v2, v3, v3, v2
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-64-LABEL: no_crash_bitcast:
+; P8-AIX-64:       # %bb.0: # %entry
+; P8-AIX-64-NEXT:    mtfprwz f0, r3
+; P8-AIX-64-NEXT:    xxmrghw v2, vs0, vs0
+; P8-AIX-64-NEXT:    blr
+;
+; P8-AIX-32-LABEL: no_crash_bitcast:
+; P8-AIX-32:       # %bb.0: # %entry
+; P8-AIX-32-NEXT:    lwz r4, L..C1(r2) # %const.0
+; P8-AIX-32-NEXT:    stw r3, -16(r1)
+; P8-AIX-32-NEXT:    addi r3, r1, -16
+; P8-AIX-32-NEXT:    lxvw4x v3, 0, r3
+; P8-AIX-32-NEXT:    lxvw4x v2, 0, r4
+; P8-AIX-32-NEXT:    vperm v2, v3, v3, v2
+; P8-AIX-32-NEXT:    blr
 entry:
   %cast = bitcast i32 %a to <4 x i8>
   %ret = shufflevector <4 x i8> %cast, <4 x i8> undef, <16 x i32> <i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef>
@@ -632,6 +768,24 @@ define dso_local <4 x i32> @replace_undefs_in_splat(<4 x i32> %a) local_unnamed_
 ; CHECK-P7-NEXT:    lvx v4, 0, r3
 ; CHECK-P7-NEXT:    vperm v2, v4, v2, v3
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-64-LABEL: replace_undefs_in_splat:
+; P8-AIX-64:       # %bb.0: # %entry
+; P8-AIX-64-NEXT:    ld r3, L..C1(r2) # %const.0
+; P8-AIX-64-NEXT:    ld r4, L..C2(r2) # %const.1
+; P8-AIX-64-NEXT:    lxvw4x v3, 0, r3
+; P8-AIX-64-NEXT:    lxvw4x v4, 0, r4
+; P8-AIX-64-NEXT:    vperm v2, v2, v4, v3
+; P8-AIX-64-NEXT:    blr
+;
+; P8-AIX-32-LABEL: replace_undefs_in_splat:
+; P8-AIX-32:       # %bb.0: # %entry
+; P8-AIX-32-NEXT:    lwz r3, L..C2(r2) # %const.0
+; P8-AIX-32-NEXT:    lwz r4, L..C3(r2) # %const.1
+; P8-AIX-32-NEXT:    lxvw4x v3, 0, r3
+; P8-AIX-32-NEXT:    lxvw4x v4, 0, r4
+; P8-AIX-32-NEXT:    vperm v2, v2, v4, v3
+; P8-AIX-32-NEXT:    blr
 entry:
   %vecins1 = shufflevector <4 x i32> %a, <4 x i32> <i32 undef, i32 566, i32 undef, i32 566>, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
   ret <4 x i32> %vecins1
@@ -683,6 +837,30 @@ define dso_local <16 x i8> @no_RAUW_in_combine_during_legalize(i32* nocapture re
 ; CHECK-P7-NEXT:    xxspltd v2, f0, 0
 ; CHECK-P7-NEXT:    vmrglb v2, v3, v2
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-64-LABEL: no_RAUW_in_combine_during_legalize:
+; P8-AIX-64:       # %bb.0: # %entry
+; P8-AIX-64-NEXT:    sldi r4, r4, 2
+; P8-AIX-64-NEXT:    xxlxor v3, v3, v3
+; P8-AIX-64-NEXT:    lxsiwzx v2, r3, r4
+; P8-AIX-64-NEXT:    vmrghb v2, v2, v3
+; P8-AIX-64-NEXT:    blr
+;
+; P8-AIX-32-LABEL: no_RAUW_in_combine_during_legalize:
+; P8-AIX-32:       # %bb.0: # %entry
+; P8-AIX-32-NEXT:    slwi r4, r4, 2
+; P8-AIX-32-NEXT:    xxlxor v3, v3, v3
+; P8-AIX-32-NEXT:    lwzx r3, r3, r4
+; P8-AIX-32-NEXT:    li r4, 0
+; P8-AIX-32-NEXT:    stw r4, -32(r1)
+; P8-AIX-32-NEXT:    addi r4, r1, -16
+; P8-AIX-32-NEXT:    stw r3, -16(r1)
+; P8-AIX-32-NEXT:    addi r3, r1, -32
+; P8-AIX-32-NEXT:    lxvw4x vs0, 0, r3
+; P8-AIX-32-NEXT:    lxvw4x vs1, 0, r4
+; P8-AIX-32-NEXT:    xxmrghw v2, vs0, vs1
+; P8-AIX-32-NEXT:    vmrghb v2, v2, v3
+; P8-AIX-32-NEXT:    blr
 entry:
   %idx.ext = sext i32 %offset to i64
   %add.ptr = getelementptr inbounds i32, i32* %ptr, i64 %idx.ext
@@ -731,6 +909,19 @@ define dso_local <4 x i32> @testSplat4Low(<8 x i8>* nocapture readonly %ptr) loc
 ; CHECK-P7-NEXT:    xxswapd v2, vs0
 ; CHECK-P7-NEXT:    xxspltw v2, v2, 2
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-64-LABEL: testSplat4Low:
+; P8-AIX-64:       # %bb.0: # %entry
+; P8-AIX-64-NEXT:    lfdx f0, 0, r3
+; P8-AIX-64-NEXT:    xxspltw v2, vs0, 1
+; P8-AIX-64-NEXT:    blr
+;
+; P8-AIX-32-LABEL: testSplat4Low:
+; P8-AIX-32:       # %bb.0: # %entry
+; P8-AIX-32-NEXT:    addi r3, r3, 4
+; P8-AIX-32-NEXT:    lfiwzx f0, 0, r3
+; P8-AIX-32-NEXT:    xxspltw v2, vs0, 1
+; P8-AIX-32-NEXT:    blr
 entry:
   %0 = load <8 x i8>, <8 x i8>* %ptr, align 8
   %vecinit18 = shufflevector <8 x i8> %0, <8 x i8> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
@@ -774,6 +965,19 @@ define dso_local <4 x i32> @testSplat4hi(<8 x i8>* nocapture readonly %ptr) loca
 ; CHECK-P7-NEXT:    xxswapd v2, vs0
 ; CHECK-P7-NEXT:    xxspltw v2, v2, 3
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-64-LABEL: testSplat4hi:
+; P8-AIX-64:       # %bb.0: # %entry
+; P8-AIX-64-NEXT:    lfdx f0, 0, r3
+; P8-AIX-64-NEXT:    xxspltw v2, vs0, 0
+; P8-AIX-64-NEXT:    blr
+;
+; P8-AIX-32-LABEL: testSplat4hi:
+; P8-AIX-32:       # %bb.0: # %entry
+; P8-AIX-32-NEXT:    lwz r3, 0(r3)
+; P8-AIX-32-NEXT:    mtfprwz f0, r3
+; P8-AIX-32-NEXT:    xxspltw v2, vs0, 1
+; P8-AIX-32-NEXT:    blr
 entry:
   %0 = load <8 x i8>, <8 x i8>* %ptr, align 8
   %vecinit22 = shufflevector <8 x i8> %0, <8 x i8> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
@@ -814,6 +1018,27 @@ define dso_local <2 x i64> @testSplat8(<8 x i8>* nocapture readonly %ptr) local_
 ; CHECK-P7:       # %bb.0: # %entry
 ; CHECK-P7-NEXT:    lxvdsx v2, 0, r3
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-64-LABEL: testSplat8:
+; P8-AIX-64:       # %bb.0: # %entry
+; P8-AIX-64-NEXT:    lxvdsx v2, 0, r3
+; P8-AIX-64-NEXT:    blr
+;
+; P8-AIX-32-LABEL: testSplat8:
+; P8-AIX-32:       # %bb.0: # %entry
+; P8-AIX-32-NEXT:    lwz r5, 4(r3)
+; P8-AIX-32-NEXT:    lwz r4, L..C4(r2) # %const.0
+; P8-AIX-32-NEXT:    stw r5, -32(r1)
+; P8-AIX-32-NEXT:    lwz r3, 0(r3)
+; P8-AIX-32-NEXT:    lxvw4x v2, 0, r4
+; P8-AIX-32-NEXT:    addi r4, r1, -16
+; P8-AIX-32-NEXT:    stw r3, -16(r1)
+; P8-AIX-32-NEXT:    addi r3, r1, -32
+; P8-AIX-32-NEXT:    lxvw4x v3, 0, r3
+; P8-AIX-32-NEXT:    lxvw4x v4, 0, r4
+; P8-AIX-32-NEXT:    vperm v2, v4, v3, v2
+; P8-AIX-32-NEXT:    xxmrghd v2, v2, v2
+; P8-AIX-32-NEXT:    blr
 entry:
   %0 = load <8 x i8>, <8 x i8>* %ptr, align 8
   %vecinit30 = shufflevector <8 x i8> %0, <8 x i8> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
@@ -850,6 +1075,26 @@ define <2 x i64> @testSplati64_0(<1 x i64>* nocapture readonly %ptr) #0 {
 ; CHECK-P7:       # %bb.0: # %entry
 ; CHECK-P7-NEXT:    lxvdsx v2, 0, r3
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-64-LABEL: testSplati64_0:
+; P8-AIX-64:       # %bb.0: # %entry
+; P8-AIX-64-NEXT:    lxvdsx v2, 0, r3
+; P8-AIX-64-NEXT:    blr
+;
+; P8-AIX-32-LABEL: testSplati64_0:
+; P8-AIX-32:       # %bb.0: # %entry
+; P8-AIX-32-NEXT:    lwz r4, L..C5(r2) # %const.0
+; P8-AIX-32-NEXT:    lwz r5, 4(r3)
+; P8-AIX-32-NEXT:    lwz r3, 0(r3)
+; P8-AIX-32-NEXT:    stw r5, -16(r1)
+; P8-AIX-32-NEXT:    stw r3, -32(r1)
+; P8-AIX-32-NEXT:    addi r3, r1, -16
+; P8-AIX-32-NEXT:    lxvw4x v2, 0, r4
+; P8-AIX-32-NEXT:    addi r4, r1, -32
+; P8-AIX-32-NEXT:    lxvw4x v3, 0, r3
+; P8-AIX-32-NEXT:    lxvw4x v4, 0, r4
+; P8-AIX-32-NEXT:    vperm v2, v4, v3, v2
+; P8-AIX-32-NEXT:    blr
 entry:
   %0 = load <1 x i64>, <1 x i64>* %ptr, align 8
   %1 = shufflevector <1 x i64> %0, <1 x i64> undef, <2 x i32> <i32 0, i32 0>
@@ -894,6 +1139,12 @@ define <2 x i64> @testSplati64_1(<2 x i64>* nocapture readonly %ptr) #0 {
 ; CHECK-P7-NEXT:    lxvd2x vs0, 0, r3
 ; CHECK-P7-NEXT:    xxspltd v2, vs0, 1
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-LABEL: testSplati64_1:
+; P8-AIX:       # %bb.0: # %entry
+; P8-AIX-NEXT:    addi r3, r3, 8
+; P8-AIX-NEXT:    lxvdsx v2, 0, r3
+; P8-AIX-NEXT:    blr
 entry:
   %0 = load <2 x i64>, <2 x i64>* %ptr, align 8
   %1 = shufflevector <2 x i64> %0, <2 x i64> undef, <2 x i32> <i32 1, i32 1>
@@ -941,6 +1192,14 @@ define dso_local void @testByteSplat() #0 {
 ; CHECK-P7-NEXT:    vspltb v2, v2, 15
 ; CHECK-P7-NEXT:    stvx v2, 0, r3
 ; CHECK-P7-NEXT:    blr
+;
+; P8-AIX-LABEL: testByteSplat:
+; P8-AIX:       # %bb.0: # %entry
+; P8-AIX-NEXT:    lbzx r3, 0, r3
+; P8-AIX-NEXT:    mtvsrwz v2, r3
+; P8-AIX-NEXT:    vspltb v2, v2, 7
+; P8-AIX-NEXT:    stxvw4x v2, 0, r3
+; P8-AIX-NEXT:    blr
 entry:
   %0 = load i8, i8* undef, align 1
   %splat.splatinsert.i = insertelement <16 x i8> poison, i8 %0, i32 0
diff --git a/llvm/test/CodeGen/PowerPC/float-vector-gather.ll b/llvm/test/CodeGen/PowerPC/float-vector-gather.ll
index 672a8f3e82aa..0b3e75e1c370 100644
--- a/llvm/test/CodeGen/PowerPC/float-vector-gather.ll
+++ b/llvm/test/CodeGen/PowerPC/float-vector-gather.ll
@@ -6,6 +6,12 @@
 ; RUN: llc -verify-machineinstrs -mcpu=pwr9 -ppc-vsr-nums-as-vr \
 ; RUN: -ppc-asm-full-reg-names -mtriple=powerpc64-unknown-linux-gnu < %s \
 ; RUN: | FileCheck %s -check-prefix=CHECK-BE
+; RUN: llc -verify-machineinstrs -mcpu=pwr9 -ppc-vsr-nums-as-vr \
+; RUN: -ppc-asm-full-reg-names -mtriple=powerpc64-ibm-aix-xcoff < %s \
+; RUN: | FileCheck %s -check-prefix=CHECK-BE
+; RUN: llc -verify-machineinstrs -mcpu=pwr9 -ppc-vsr-nums-as-vr \
+; RUN: -ppc-asm-full-reg-names -mtriple=powerpc-ibm-aix-xcoff < %s \
+; RUN: | FileCheck %s -check-prefix=CHECK-BE-AIX-32
 define dso_local <4 x float> @vector_gatherf(float* nocapture readonly %a,
 float* nocapture readonly %b, float* nocapture readonly %c,
 float* nocapture readonly %d) {
@@ -34,6 +40,21 @@ float* nocapture readonly %d) {
 ; CHECK-BE-DAG:    xxmrghw vs[[REG4:[0-9]+]], vs[[REG2]], vs[[REG3]]
 ; CHECK-BE-NEXT:   xxmrgld v[[REG:[0-9]+]], vs[[REG0]], vs[[REG4]]
 ; CHECK-BE-NEXT:   blr
+
+; CHECK-BE-AIX-32-LABEL: vector_gatherf:
+; CHECK-BE-AIX-32-LABEL: # %bb.0: # %entry
+; CHECK-BE-AIX-32-DAG: lfs f[[REG0:[0-9]+]]
+; CHECK-BE-AIX-32-DAG: lfs f[[REG1:[0-9]+]]
+; CHECK-BE-AIX-32-DAG: lfs f[[REG2:[0-9]+]]
+; CHECK-BE-AIX-32-DAG: lfs f[[REG3:[0-9]+]]
+; CHECK-BE-AIX-32-DAG: xscvdpspn v[[VREG0:[0-9]+]], f[[REG0]]
+; CHECK-BE-AIX-32-DAG: xscvdpspn v[[VREG1:[0-9]+]], f[[REG1]]
+; CHECK-BE-AIX-32-DAG: xscvdpspn v[[VREG2:[0-9]+]], f[[REG2]]
+; CHECK-BE-AIX-32-DAG: xscvdpspn v[[VREG0:[0-9]+]], f[[REG3]]
+; CHECK-BE-AIX-32-DAG: vmrgow v[[VREG1]], v[[VREG0]], v[[VREG1]]
+; CHECK-BE-AIX-32-DAG: vmrgow v[[VREG0]], v[[VREG2]], v[[VREG0]]
+; CHECK-BE-AIX-32-NEXT: xxmrghd v[[VREG1]], v[[VREG0]], v[[VREG1]]
+; CHECK-BE-AIX-32-NEXT: blr
 entry:
   %0 = load float, float* %a, align 4
   %vecinit = insertelement <4 x float> undef, float %0, i32 0
diff --git a/llvm/test/CodeGen/PowerPC/load-v4i8-improved.ll b/llvm/test/CodeGen/PowerPC/load-v4i8-improved.ll
index 42b449f22776..816026379224 100644
--- a/llvm/test/CodeGen/PowerPC/load-v4i8-improved.ll
+++ b/llvm/test/CodeGen/PowerPC/load-v4i8-improved.ll
@@ -1,9 +1,17 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -verify-machineinstrs -mcpu=pwr8 -mtriple=powerpc64le-unknown-linux-gnu < %s \
 ; RUN:   -ppc-vsr-nums-as-vr -ppc-asm-full-reg-names | FileCheck --check-prefix=CHECK-LE \
 ; RUN:   -implicit-check-not vmrg -implicit-check-not=vperm %s
 ; RUN: llc -verify-machineinstrs -mcpu=pwr8 -mtriple=powerpc64-unknown-linux-gnu < %s \
 ; RUN:   -ppc-vsr-nums-as-vr -ppc-asm-full-reg-names | FileCheck \
 ; RUN:   -implicit-check-not vmrg -implicit-check-not=vperm %s
+; RUN: llc -verify-machineinstrs -mcpu=pwr8 -mtriple=powerpc64-ibm-aix-xcoff < %s \
+; RUN:   -ppc-vsr-nums-as-vr -ppc-asm-full-reg-names | FileCheck \
+; RUN:   -implicit-check-not vmrg -implicit-check-not=vperm %s
+; RUN: llc -verify-machineinstrs -mcpu=pwr8 -mtriple=powerpc-ibm-aix-xcoff < %s \
+; RUN:   -ppc-vsr-nums-as-vr -ppc-asm-full-reg-names | FileCheck \
+; RUN:   --check-prefix=CHECK-AIX-32 -implicit-check-not vmrg \
+; RUN:   -implicit-check-not=vperm %s
 
 define <16 x i8> @test(i32* %s, i32* %t) {
 ; CHECK-LE-LABEL: test:
@@ -11,12 +19,20 @@ define <16 x i8> @test(i32* %s, i32* %t) {
 ; CHECK-LE-NEXT:    lfiwzx f0, 0, r3
 ; CHECK-LE-NEXT:    xxspltw v2, vs0, 1
 ; CHECK-LE-NEXT:    blr
-
+;
 ; CHECK-LABEL: test:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    lfiwzx f0, 0, r3
 ; CHECK-NEXT:    xxspltw v2, vs0, 1
 ; CHECK-NEXT:    blr
+;
+; CHECK-AIX-32-LABEL: test:
+; CHECK-AIX-32:       # %bb.0: # %entry
+; CHECK-AIX-32-NEXT:    lwz r3, 0(r3)
+; CHECK-AIX-32-NEXT:    mtfprwz f0, r3
+; CHECK-AIX-32-NEXT:    xxspltw v2, vs0, 1
+; CHECK-AIX-32-NEXT:    blr
+
 entry:
   %0 = bitcast i32* %s to <4 x i8>*
   %1 = load <4 x i8>, <4 x i8>* %0, align 4
diff --git a/llvm/test/CodeGen/PowerPC/pre-inc-disable.ll b/llvm/test/CodeGen/PowerPC/pre-inc-disable.ll
index 5ab2da486963..f3959c3c8ec9 100644
--- a/llvm/test/CodeGen/PowerPC/pre-inc-disable.ll
+++ b/llvm/test/CodeGen/PowerPC/pre-inc-disable.ll
@@ -10,6 +10,10 @@
 ; RUN:     -mtriple=powerpc64-ibm-aix-xcoff -vec-extabi \
 ; RUN:     < %s | FileCheck %s --check-prefix=P9BE
 
+; RUN: llc -mcpu=pwr9 -O3 -verify-machineinstrs -ppc-vsr-nums-as-vr \
+; RUN:     -mtriple=powerpc-ibm-aix-xcoff -vec-extabi \
+; RUN:     < %s | FileCheck %s --check-prefix=P9BE-32
+
 define void @test64(i8* nocapture readonly %pix2, i32 signext %i_pix2) {
 ; CHECK-LABEL: test64:
 ; CHECK-NOT: ldux
@@ -21,6 +25,11 @@ define void @test64(i8* nocapture readonly %pix2, i32 signext %i_pix2) {
 ; P9BE-NOT: mtvsrd
 ; P9BE: lxsdx [[REG:[0-9]+]]
 ; P9BE: vperm {{[0-9]+}}, {{[0-9]+}}, [[REG]]
+; P9BE-32-LABEL: test64:
+; P9BE-32: lwzux [[REG1:[0-9]+]]
+; P9BE-32: mtfprwz [[REG2:[0-9]+]], [[REG1]]
+; P9BE-32: xxinsertw [[REG3:[0-9]+]], [[REG2]]
+; P9BE-32: vperm {{[0-9]+}}, {{[0-9]+}}, [[REG3]]
 entry:
   %idx.ext63 = sext i32 %i_pix2 to i64
   %add.ptr64 = getelementptr inbounds i8, i8* %pix2, i64 %idx.ext63
@@ -56,6 +65,10 @@ define void @test32(i8* nocapture readonly %pix2, i32 signext %i_pix2) {
 ; P9BE-NOT: mtvsrwz
 ; P9BE: lxsiwzx [[REG:[0-9]+]]
 ; P9BE: vperm {{[0-9]+}}, {{[0-9]+}}, [[REG]]
+; P9BE-32-LABEL: test32:
+; P9BE-32: lwzx [[REG1:[0-9]+]]
+; P9BE-32: mtvsrwz [[REG2:[0-9]+]], [[REG1]]
+; P9BE-32: vperm {{[0-9]+}}, {{[0-9]+}}, [[REG2]]
 entry:
   %idx.ext63 = sext i32 %i_pix2 to i64
   %add.ptr64 = getelementptr inbounds i8, i8* %pix2, i64 %idx.ext63
@@ -89,6 +102,9 @@ define void @test16(i16* nocapture readonly %sums, i32 signext %delta, i32 signe
 ; P9BE-NOT: lhzux
 ; P9BE: lxsihzx [[REG:[0-9]+]]
 ; P9BE: vperm {{[0-9]+}}, {{[0-9]+}}, [[REG]]
+; P9BE-32-LABEL: test16:
+; P9BE-32: lhzux [[REG1:[0-9]+]]
+; P9BE-32: vmrghh {{[0-9]+}}, {{[0-9]+}}, [[REG1]]
 entry:
   %idxprom = sext i32 %delta to i64
   %add14 = add nsw i32 %delta, 8
@@ -130,6 +146,9 @@ define void @test8(i8* nocapture readonly %sums, i32 signext %delta, i32 signext
 ; P9BE-NOT: lbzux
 ; P9BE: lxsibzx [[REG:[0-9]+]]
 ; P9BE: vperm {{[0-9]+}}, {{[0-9]+}}, [[REG]]
+; P9BE-32-LABEL: test8:
+; P9BE-32: lxsibzx [[REG:[0-9]+]]
+; P9BE-32: vperm {{[0-9]+}}, {{[0-9]+}}, [[REG]]
 entry:
   %idxprom = sext i32 %delta to i64
   %add14 = add nsw i32 %delta, 8
diff --git a/llvm/test/CodeGen/PowerPC/reduce_scalarization.ll b/llvm/test/CodeGen/PowerPC/reduce_scalarization.ll
index 5fdd21e9a0b6..d7883b540d89 100644
--- a/llvm/test/CodeGen/PowerPC/reduce_scalarization.ll
+++ b/llvm/test/CodeGen/PowerPC/reduce_scalarization.ll
@@ -11,6 +11,12 @@
 ; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-unknown \
 ; RUN:     -mcpu=pwr10 -ppc-asm-full-reg-names -ppc-vsr-nums-as-vr \
 ; RUN:     < %s | FileCheck %s --check-prefixes=CHECK,CHECK-P10-BE
+; RUN: llc -mcpu=pwr8 -verify-machineinstrs -ppc-vsr-nums-as-vr \
+; RUN:   -ppc-asm-full-reg-names -mtriple=powerpc64-ibm-aix-xcoff < %s | \
+; RUN: FileCheck %s --check-prefix=AIX-64
+; RUN: llc -mcpu=pwr8 -verify-machineinstrs -ppc-vsr-nums-as-vr \
+; RUN:   -ppc-asm-full-reg-names -mtriple=powerpc-ibm-aix-xcoff < %s | \
+; RUN: FileCheck %s --check-prefix=AIX-32
 
 ; Function Attrs: norecurse nounwind readonly
 define dso_local <2 x double> @test1(<2 x float>* nocapture readonly %Ptr) {
@@ -20,6 +26,20 @@ define dso_local <2 x double> @test1(<2 x float>* nocapture readonly %Ptr) {
 ; CHECK-NEXT:    xxmrghw vs0, vs0, vs0
 ; CHECK-NEXT:    xvcvspdp v2, vs0
 ; CHECK-NEXT:    blr
+;
+; AIX-64-LABEL: test1:
+; AIX-64:       # %bb.0: # %entry
+; AIX-64-NEXT:    lfdx f0, 0, r3
+; AIX-64-NEXT:    xxmrghw vs0, vs0, vs0
+; AIX-64-NEXT:    xvcvspdp v2, vs0
+; AIX-64-NEXT:    blr
+;
+; AIX-32-LABEL: test1:
+; AIX-32:       # %bb.0: # %entry
+; AIX-32-NEXT:    lfs f0, 4(r3)
+; AIX-32-NEXT:    lfs f1, 0(r3)
+; AIX-32-NEXT:    xxmrghd v2, vs1, vs0
+; AIX-32-NEXT:    blr
 entry:
   %0 = load <2 x float>, <2 x float>* %Ptr, align 8
   %1 = fpext <2 x float> %0 to <2 x double>
@@ -36,6 +56,36 @@ define dso_local <2 x double> @test2(<2 x float>* nocapture readonly %a, <2 x fl
 ; CHECK-NEXT:    xxmrghw vs0, vs0, vs0
 ; CHECK-NEXT:    xvcvspdp v2, vs0
 ; CHECK-NEXT:    blr
+;
+; AIX-64-LABEL: test2:
+; AIX-64:       # %bb.0: # %entry
+; AIX-64-NEXT:    lfdx f0, 0, r3
+; AIX-64-NEXT:    lfdx f1, 0, r4
+; AIX-64-NEXT:    xvsubsp vs0, vs0, vs1
+; AIX-64-NEXT:    xxmrghw vs0, vs0, vs0
+; AIX-64-NEXT:    xvcvspdp v2, vs0
+; AIX-64-NEXT:    blr
+;
+; AIX-32-LABEL: test2:
+; AIX-32:       # %bb.0: # %entry
+; AIX-32-NEXT:    lfs f0, 4(r3)
+; AIX-32-NEXT:    lfs f1, 0(r3)
+; AIX-32-NEXT:    lwz r5, L..C0(r2) # %const.0
+; AIX-32-NEXT:    lfs f2, 4(r4)
+; AIX-32-NEXT:    xscvdpspn v2, f0
+; AIX-32-NEXT:    lfs f0, 0(r4)
+; AIX-32-NEXT:    lxvw4x v0, 0, r5
+; AIX-32-NEXT:    xscvdpspn v3, f1
+; AIX-32-NEXT:    xscvdpspn v4, f2
+; AIX-32-NEXT:    xscvdpspn v5, f0
+; AIX-32-NEXT:    vperm v2, v3, v2, v0
+; AIX-32-NEXT:    vperm v3, v5, v4, v0
+; AIX-32-NEXT:    xvsubsp vs0, v2, v3
+; AIX-32-NEXT:    xxsldwi vs1, vs0, vs0, 1
+; AIX-32-NEXT:    xscvspdpn f0, vs0
+; AIX-32-NEXT:    xscvspdpn f1, vs1
+; AIX-32-NEXT:    xxmrghd v2, vs0, vs1
+; AIX-32-NEXT:    blr
 entry:
   %0 = load <2 x float>, <2 x float>* %a, align 8
   %1 = load <2 x float>, <2 x float>* %b, align 8
@@ -55,6 +105,36 @@ define dso_local <2 x double> @test3(<2 x float>* nocapture readonly %a, <2 x fl
 ; CHECK-NEXT:    xxmrghw vs0, vs0, vs0
 ; CHECK-NEXT:    xvcvspdp v2, vs0
 ; CHECK-NEXT:    blr
+;
+; AIX-64-LABEL: test3:
+; AIX-64:       # %bb.0: # %entry
+; AIX-64-NEXT:    lfdx f0, 0, r3
+; AIX-64-NEXT:    lfdx f1, 0, r4
+; AIX-64-NEXT:    xvaddsp vs0, vs0, vs1
+; AIX-64-NEXT:    xxmrghw vs0, vs0, vs0
+; AIX-64-NEXT:    xvcvspdp v2, vs0
+; AIX-64-NEXT:    blr
+;
+; AIX-32-LABEL: test3:
+; AIX-32:       # %bb.0: # %entry
+; AIX-32-NEXT:    lfs f0, 4(r3)
+; AIX-32-NEXT:    lfs f1, 0(r3)
+; AIX-32-NEXT:    lwz r5, L..C1(r2) # %const.0
+; AIX-32-NEXT:    lfs f2, 4(r4)
+; AIX-32-NEXT:    xscvdpspn v2, f0
+; AIX-32-NEXT:    lfs f0, 0(r4)
+; AIX-32-NEXT:    lxvw4x v0, 0, r5
+; AIX-32-NEXT:    xscvdpspn v3, f1
+; AIX-32-NEXT:    xscvdpspn v4, f2
+; AIX-32-NEXT:    xscvdpspn v5, f0
+; AIX-32-NEXT:    vperm v2, v3, v2, v0
+; AIX-32-NEXT:    vperm v3, v5, v4, v0
+; AIX-32-NEXT:    xvaddsp vs0, v2, v3
+; AIX-32-NEXT:    xxsldwi vs1, vs0, vs0, 1
+; AIX-32-NEXT:    xscvspdpn f0, vs0
+; AIX-32-NEXT:    xscvspdpn f1, vs1
+; AIX-32-NEXT:    xxmrghd v2, vs0, vs1
+; AIX-32-NEXT:    blr
 entry:
   %0 = load <2 x float>, <2 x float>* %a, align 8
   %1 = load <2 x float>, <2 x float>* %b, align 8
@@ -74,6 +154,36 @@ define dso_local <2 x double> @test4(<2 x float>* nocapture readonly %a, <2 x fl
 ; CHECK-NEXT:    xxmrghw vs0, vs0, vs0
 ; CHECK-NEXT:    xvcvspdp v2, vs0
 ; CHECK-NEXT:    blr
+;
+; AIX-64-LABEL: test4:
+; AIX-64:       # %bb.0: # %entry
+; AIX-64-NEXT:    lfdx f0, 0, r3
+; AIX-64-NEXT:    lfdx f1, 0, r4
+; AIX-64-NEXT:    xvmulsp vs0, vs0, vs1
+; AIX-64-NEXT:    xxmrghw vs0, vs0, vs0
+; AIX-64-NEXT:    xvcvspdp v2, vs0
+; AIX-64-NEXT:    blr
+;
+; AIX-32-LABEL: test4:
+; AIX-32:       # %bb.0: # %entry
+; AIX-32-NEXT:    lfs f0, 4(r3)
+; AIX-32-NEXT:    lfs f1, 0(r3)
+; AIX-32-NEXT:    lwz r5, L..C2(r2) # %const.0
+; AIX-32-NEXT:    lfs f2, 4(r4)
+; AIX-32-NEXT:    xscvdpspn v2, f0
+; AIX-32-NEXT:    lfs f0, 0(r4)
+; AIX-32-NEXT:    lxvw4x v0, 0, r5
+; AIX-32-NEXT:    xscvdpspn v3, f1
+; AIX-32-NEXT:    xscvdpspn v4, f2
+; AIX-32-NEXT:    xscvdpspn v5, f0
+; AIX-32-NEXT:    vperm v2, v3, v2, v0
+; AIX-32-NEXT:    vperm v3, v5, v4, v0
+; AIX-32-NEXT:    xvmulsp vs0, v2, v3
+; AIX-32-NEXT:    xxsldwi vs1, vs0, vs0, 1
+; AIX-32-NEXT:    xscvspdpn f0, vs0
+; AIX-32-NEXT:    xscvspdpn f1, vs1
+; AIX-32-NEXT:    xxmrghd v2, vs0, vs1
+; AIX-32-NEXT:    blr
 entry:
   %0 = load <2 x float>, <2 x float>* %a, align 8
   %1 = load <2 x float>, <2 x float>* %b, align 8
@@ -102,6 +212,24 @@ define dso_local <2 x double> @test5(<2 x double> %a) {
 ; CHECK-P10-BE-NEXT:    xvcvspdp vs0, vs0
 ; CHECK-P10-BE-NEXT:    xvadddp v2, vs0, v2
 ; CHECK-P10-BE-NEXT:    blr
+;
+; AIX-64-LABEL: test5:
+; AIX-64:       # %bb.0: # %entry
+; AIX-64-NEXT:    ld r3, L..C0(r2) # @G
+; AIX-64-NEXT:    lfdx f0, 0, r3
+; AIX-64-NEXT:    xxmrghw vs0, vs0, vs0
+; AIX-64-NEXT:    xvcvspdp vs0, vs0
+; AIX-64-NEXT:    xvadddp v2, vs0, v2
+; AIX-64-NEXT:    blr
+;
+; AIX-32-LABEL: test5:
+; AIX-32:       # %bb.0: # %entry
+; AIX-32-NEXT:    lwz r3, L..C3(r2) # @G
+; AIX-32-NEXT:    lfs f0, 4(r3)
+; AIX-32-NEXT:    lfs f1, 0(r3)
+; AIX-32-NEXT:    xxmrghd vs0, vs1, vs0
+; AIX-32-NEXT:    xvadddp v2, vs0, v2
+; AIX-32-NEXT:    blr
 entry:
   %0 = load <2 x float>, <2 x float>* @G, align 8
   %1 = fpext <2 x float> %0 to <2 x double>
@@ -144,6 +272,44 @@ define dso_local i32 @test6() #0 {
 ; CHECK-P10-BE-NEXT:    bc 4, gt, .LBB5_2
 ; CHECK-P10-BE-NEXT:  # %bb.1: # %bb8
 ; CHECK-P10-BE-NEXT:  .LBB5_2: # %bb7
+;
+; AIX-64-LABEL: test6:
+; AIX-64:       # %bb.0: # %bb
+; AIX-64-NEXT:    ld r3, L..C1(r2) # @Glob1
+; AIX-64-NEXT:    lis r4, 8
+; AIX-64-NEXT:    xxlxor vs1, vs1, vs1
+; AIX-64-NEXT:    ori r4, r4, 38248
+; AIX-64-NEXT:    lfdx f0, r3, r4
+; AIX-64-NEXT:    xxmrghw vs0, vs0, vs0
+; AIX-64-NEXT:    xvcvspdp vs0, vs0
+; AIX-64-NEXT:    xvcmpeqdp v2, vs1, vs0
+; AIX-64-NEXT:    xxswapd v3, v2
+; AIX-64-NEXT:    xxland vs0, v2, v3
+; AIX-64-NEXT:    mffprd r3, f0
+; AIX-64-NEXT:    andi. r3, r3, 1
+; AIX-64-NEXT:    bc 4, gt, L..BB5_2
+; AIX-64-NEXT:  # %bb.1: # %bb8
+; AIX-64-NEXT:  L..BB5_2: # %bb7
+;
+; AIX-32-LABEL: test6:
+; AIX-32:       # %bb.0: # %bb
+; AIX-32-NEXT:    lwz r3, L..C4(r2) # @Glob1
+; AIX-32-NEXT:    lis r4, 8
+; AIX-32-NEXT:    ori r4, r4, 38248
+; AIX-32-NEXT:    lfsux f0, r3, r4
+; AIX-32-NEXT:    lfs f1, 4(r3)
+; AIX-32-NEXT:    addi r3, r1, -16
+; AIX-32-NEXT:    xxmrghd vs0, vs0, vs1
+; AIX-32-NEXT:    xxlxor vs1, vs1, vs1
+; AIX-32-NEXT:    xvcmpeqdp v2, vs1, vs0
+; AIX-32-NEXT:    xxswapd v3, v2
+; AIX-32-NEXT:    xxland vs0, v2, v3
+; AIX-32-NEXT:    stxvw4x vs0, 0, r3
+; AIX-32-NEXT:    lwz r3, -12(r1)
+; AIX-32-NEXT:    andi. r3, r3, 1
+; AIX-32-NEXT:    bc 4, gt, L..BB5_2
+; AIX-32-NEXT:  # %bb.1: # %bb8
+; AIX-32-NEXT:  L..BB5_2: # %bb7
 bb:
   br label %bb1
 
diff --git a/llvm/test/CodeGen/PowerPC/scalar_vector_test_4.ll b/llvm/test/CodeGen/PowerPC/scalar_vector_test_4.ll
index b4e62073ec29..e0fd7526faaf 100644
--- a/llvm/test/CodeGen/PowerPC/scalar_vector_test_4.ll
+++ b/llvm/test/CodeGen/PowerPC/scalar_vector_test_4.ll
@@ -8,6 +8,19 @@
 ; RUN: llc -mcpu=pwr8 -verify-machineinstrs -ppc-vsr-nums-as-vr -ppc-asm-full-reg-names \
 ; RUN:    -mtriple=powerpc64-unknown-linux-gnu < %s | FileCheck %s --check-prefix=P8BE
 
+; RUN: llc -mcpu=pwr9 -verify-machineinstrs -ppc-vsr-nums-as-vr -ppc-asm-full-reg-names \
+; RUN:    -mtriple=powerpc64-ibm-aix-xcoff< %s | FileCheck %s \
+; RUN:    --check-prefixes=P9-AIX,P9-AIX-64
+; RUN: llc -mcpu=pwr9 -verify-machineinstrs -ppc-vsr-nums-as-vr -ppc-asm-full-reg-names \
+; RUN:    -mtriple=powerpc-ibm-aix-xcoff < %s | FileCheck %s \
+; RUN:    --check-prefixes=P9-AIX,P9-AIX-32
+; RUN: llc -mcpu=pwr8 -verify-machineinstrs -ppc-vsr-nums-as-vr -ppc-asm-full-reg-names \
+; RUN:    -mtriple=powerpc64-ibm-aix-xcoff < %s | FileCheck %s \
+; RUN:    --check-prefixes=P8-AIX-64
+; RUN: llc -mcpu=pwr8 -verify-machineinstrs -ppc-vsr-nums-as-vr -ppc-asm-full-reg-names \
+; RUN:    -mtriple=powerpc-ibm-aix-xcoff < %s | FileCheck %s \
+; RUN:    --check-prefixes=P8-AIX-32
+
 ; Function Attrs: norecurse nounwind readonly
 define <4 x i32> @s2v_test1(i32* nocapture readonly %int32, <4 x i32> %vec)  {
 ; P9LE-LABEL: s2v_test1:
@@ -41,6 +54,32 @@ define <4 x i32> @s2v_test1(i32* nocapture readonly %int32, <4 x i32> %vec)  {
 ; P8BE-NEXT:    lxvw4x v3, 0, r4
 ; P8BE-NEXT:    vperm v2, v4, v2, v3
 ; P8BE-NEXT:    blr
+;
+; P9-AIX-LABEL: s2v_test1:
+; P9-AIX:       # %bb.0: # %entry
+; P9-AIX-NEXT:    lwz r3, 0(r3)
+; P9-AIX-NEXT:    mtfprwz f0, r3
+; P9-AIX-NEXT:    xxinsertw v2, vs0, 0
+; P9-AIX-NEXT:    blr
+;
+; P8-AIX-64-LABEL: s2v_test1:
+; P8-AIX-64:       # %bb.0: # %entry
+; P8-AIX-64-NEXT:    ld r4, L..C0(r2) # %const.0
+; P8-AIX-64-NEXT:    lxsiwzx v3, 0, r3
+; P8-AIX-64-NEXT:    lxvw4x v4, 0, r4
+; P8-AIX-64-NEXT:    vperm v2, v3, v2, v4
+; P8-AIX-64-NEXT:    blr
+;
+; P8-AIX-32-LABEL: s2v_test1:
+; P8-AIX-32:       # %bb.0: # %entry
+; P8-AIX-32-NEXT:    lwz r3, 0(r3)
+; P8-AIX-32-NEXT:    lwz r4, L..C0(r2) # %const.0
+; P8-AIX-32-NEXT:    stw r3, -16(r1)
+; P8-AIX-32-NEXT:    addi r3, r1, -16
+; P8-AIX-32-NEXT:    lxvw4x v3, 0, r4
+; P8-AIX-32-NEXT:    lxvw4x v4, 0, r3
+; P8-AIX-32-NEXT:    vperm v2, v4, v2, v3
+; P8-AIX-32-NEXT:    blr
 entry:
   %0 = load i32, i32* %int32, align 4
   %vecins = insertelement <4 x i32> %vec, i32 %0, i32 0
@@ -82,6 +121,33 @@ define <4 x i32> @s2v_test2(i32* nocapture readonly %int32, <4 x i32> %vec)  {
 ; P8BE-NEXT:    lxvw4x v3, 0, r4
 ; P8BE-NEXT:    vperm v2, v4, v2, v3
 ; P8BE-NEXT:    blr
+;
+; P9-AIX-LABEL: s2v_test2:
+; P9-AIX:       # %bb.0: # %entry
+; P9-AIX-NEXT:    lwz r3, 4(r3)
+; P9-AIX-NEXT:    mtfprwz f0, r3
+; P9-AIX-NEXT:    xxinsertw v2, vs0, 0
+; P9-AIX-NEXT:    blr
+;
+; P8-AIX-64-LABEL: s2v_test2:
+; P8-AIX-64:       # %bb.0: # %entry
+; P8-AIX-64-NEXT:    ld r4, L..C1(r2) # %const.0
+; P8-AIX-64-NEXT:    addi r3, r3, 4
+; P8-AIX-64-NEXT:    lxsiwzx v3, 0, r3
+; P8-AIX-64-NEXT:    lxvw4x v4, 0, r4
+; P8-AIX-64-NEXT:    vperm v2, v3, v2, v4
+; P8-AIX-64-NEXT:    blr
+;
+; P8-AIX-32-LABEL: s2v_test2:
+; P8-AIX-32:       # %bb.0: # %entry
+; P8-AIX-32-NEXT:    lwz r3, 4(r3)
+; P8-AIX-32-NEXT:    lwz r4, L..C1(r2) # %const.0
+; P8-AIX-32-NEXT:    stw r3, -16(r1)
+; P8-AIX-32-NEXT:    addi r3, r1, -16
+; P8-AIX-32-NEXT:    lxvw4x v3, 0, r4
+; P8-AIX-32-NEXT:    lxvw4x v4, 0, r3
+; P8-AIX-32-NEXT:    vperm v2, v4, v2, v3
+; P8-AIX-32-NEXT:    blr
 entry:
   %arrayidx = getelementptr inbounds i32, i32* %int32, i64 1
   %0 = load i32, i32* %arrayidx, align 4
@@ -126,6 +192,43 @@ define <4 x i32> @s2v_test3(i32* nocapture readonly %int32, <4 x i32> %vec, i32
 ; P8BE-NEXT:    lxvw4x v4, 0, r4
 ; P8BE-NEXT:    vperm v2, v3, v2, v4
 ; P8BE-NEXT:    blr
+;
+; P9-AIX-64-LABEL: s2v_test3:
+; P9-AIX-64:       # %bb.0: # %entry
+; P9-AIX-64-NEXT:    sldi r4, r4, 2
+; P9-AIX-64-NEXT:    lwzx r3, r3, r4
+; P9-AIX-64-NEXT:    mtfprwz f0, r3
+; P9-AIX-64-NEXT:    xxinsertw v2, vs0, 0
+; P9-AIX-64-NEXT:    blr
+;
+; P9-AIX-32-LABEL: s2v_test3:
+; P9-AIX-32:       # %bb.0: # %entry
+; P9-AIX-32-NEXT:    slwi r4, r4, 2
+; P9-AIX-32-NEXT:    lwzx r3, r3, r4
+; P9-AIX-32-NEXT:    mtfprwz f0, r3
+; P9-AIX-32-NEXT:    xxinsertw v2, vs0, 0
+; P9-AIX-32-NEXT:    blr
+;
+; P8-AIX-64-LABEL: s2v_test3:
+; P8-AIX-64:       # %bb.0: # %entry
+; P8-AIX-64-NEXT:    ld r5, L..C2(r2) # %const.0
+; P8-AIX-64-NEXT:    sldi r4, r4, 2
+; P8-AIX-64-NEXT:    lxsiwzx v3, r3, r4
+; P8-AIX-64-NEXT:    lxvw4x v4, 0, r5
+; P8-AIX-64-NEXT:    vperm v2, v3, v2, v4
+; P8-AIX-64-NEXT:    blr
+;
+; P8-AIX-32-LABEL: s2v_test3:
+; P8-AIX-32:       # %bb.0: # %entry
+; P8-AIX-32-NEXT:    slwi r4, r4, 2
+; P8-AIX-32-NEXT:    lwzx r3, r3, r4
+; P8-AIX-32-NEXT:    lwz r4, L..C2(r2) # %const.0
+; P8-AIX-32-NEXT:    stw r3, -16(r1)
+; P8-AIX-32-NEXT:    addi r3, r1, -16
+; P8-AIX-32-NEXT:    lxvw4x v3, 0, r4
+; P8-AIX-32-NEXT:    lxvw4x v4, 0, r3
+; P8-AIX-32-NEXT:    vperm v2, v4, v2, v3
+; P8-AIX-32-NEXT:    blr
 entry:
   %idxprom = sext i32 %Idx to i64
   %arrayidx = getelementptr inbounds i32, i32* %int32, i64 %idxprom
@@ -169,6 +272,33 @@ define <4 x i32> @s2v_test4(i32* nocapture readonly %int32, <4 x i32> %vec)  {
 ; P8BE-NEXT:    lxvw4x v3, 0, r4
 ; P8BE-NEXT:    vperm v2, v4, v2, v3
 ; P8BE-NEXT:    blr
+;
+; P9-AIX-LABEL: s2v_test4:
+; P9-AIX:       # %bb.0: # %entry
+; P9-AIX-NEXT:    lwz r3, 4(r3)
+; P9-AIX-NEXT:    mtfprwz f0, r3
+; P9-AIX-NEXT:    xxinsertw v2, vs0, 0
+; P9-AIX-NEXT:    blr
+;
+; P8-AIX-64-LABEL: s2v_test4:
+; P8-AIX-64:       # %bb.0: # %entry
+; P8-AIX-64-NEXT:    ld r4, L..C3(r2) # %const.0
+; P8-AIX-64-NEXT:    addi r3, r3, 4
+; P8-AIX-64-NEXT:    lxsiwzx v3, 0, r3
+; P8-AIX-64-NEXT:    lxvw4x v4, 0, r4
+; P8-AIX-64-NEXT:    vperm v2, v3, v2, v4
+; P8-AIX-64-NEXT:    blr
+;
+; P8-AIX-32-LABEL: s2v_test4:
+; P8-AIX-32:       # %bb.0: # %entry
+; P8-AIX-32-NEXT:    lwz r3, 4(r3)
+; P8-AIX-32-NEXT:    lwz r4, L..C3(r2) # %const.0
+; P8-AIX-32-NEXT:    stw r3, -16(r1)
+; P8-AIX-32-NEXT:    addi r3, r1, -16
+; P8-AIX-32-NEXT:    lxvw4x v3, 0, r4
+; P8-AIX-32-NEXT:    lxvw4x v4, 0, r3
+; P8-AIX-32-NEXT:    vperm v2, v4, v2, v3
+; P8-AIX-32-NEXT:    blr
 entry:
   %arrayidx = getelementptr inbounds i32, i32* %int32, i64 1
   %0 = load i32, i32* %arrayidx, align 4
@@ -209,6 +339,32 @@ define <4 x i32> @s2v_test5(<4 x i32> %vec, i32* nocapture readonly %ptr1)  {
 ; P8BE-NEXT:    lxvw4x v3, 0, r3
 ; P8BE-NEXT:    vperm v2, v4, v2, v3
 ; P8BE-NEXT:    blr
+;
+; P9-AIX-LABEL: s2v_test5:
+; P9-AIX:       # %bb.0: # %entry
+; P9-AIX-NEXT:    lwz r3, 0(r3)
+; P9-AIX-NEXT:    mtfprwz f0, r3
+; P9-AIX-NEXT:    xxinsertw v2, vs0, 0
+; P9-AIX-NEXT:    blr
+;
+; P8-AIX-64-LABEL: s2v_test5:
+; P8-AIX-64:       # %bb.0: # %entry
+; P8-AIX-64-NEXT:    ld r4, L..C4(r2) # %const.0
+; P8-AIX-64-NEXT:    lxsiwzx v3, 0, r3
+; P8-AIX-64-NEXT:    lxvw4x v4, 0, r4
+; P8-AIX-64-NEXT:    vperm v2, v3, v2, v4
+; P8-AIX-64-NEXT:    blr
+;
+; P8-AIX-32-LABEL: s2v_test5:
+; P8-AIX-32:       # %bb.0: # %entry
+; P8-AIX-32-NEXT:    lwz r3, 0(r3)
+; P8-AIX-32-NEXT:    lwz r4, L..C4(r2) # %const.0
+; P8-AIX-32-NEXT:    stw r3, -16(r1)
+; P8-AIX-32-NEXT:    addi r3, r1, -16
+; P8-AIX-32-NEXT:    lxvw4x v3, 0, r4
+; P8-AIX-32-NEXT:    lxvw4x v4, 0, r3
+; P8-AIX-32-NEXT:    vperm v2, v4, v2, v3
+; P8-AIX-32-NEXT:    blr
 entry:
   %0 = load i32, i32* %ptr1, align 4
   %vecins = insertelement <4 x i32> %vec, i32 %0, i32 0
@@ -248,6 +404,30 @@ define <4 x float> @s2v_test_f1(float* nocapture readonly %f64, <4 x float> %vec
 ; P8BE-NEXT:    lxvw4x v3, 0, r4
 ; P8BE-NEXT:    vperm v2, v4, v2, v3
 ; P8BE-NEXT:    blr
+;
+; P9-AIX-LABEL: s2v_test_f1:
+; P9-AIX:       # %bb.0: # %entry
+; P9-AIX-NEXT:    lwz r3, 0(r3)
+; P9-AIX-NEXT:    mtfprwz f0, r3
+; P9-AIX-NEXT:    xxinsertw v2, vs0, 0
+; P9-AIX-NEXT:    blr
+;
+; P8-AIX-64-LABEL: s2v_test_f1:
+; P8-AIX-64:       # %bb.0: # %entry
+; P8-AIX-64-NEXT:    ld r4, L..C5(r2) # %const.0
+; P8-AIX-64-NEXT:    lxsiwzx v3, 0, r3
+; P8-AIX-64-NEXT:    lxvw4x v4, 0, r4
+; P8-AIX-64-NEXT:    vperm v2, v3, v2, v4
+; P8-AIX-64-NEXT:    blr
+;
+; P8-AIX-32-LABEL: s2v_test_f1:
+; P8-AIX-32:       # %bb.0: # %entry
+; P8-AIX-32-NEXT:    lfs f0, 0(r3)
+; P8-AIX-32-NEXT:    lwz r4, L..C5(r2) # %const.0
+; P8-AIX-32-NEXT:    xscvdpspn v3, f0
+; P8-AIX-32-NEXT:    lxvw4x v4, 0, r4
+; P8-AIX-32-NEXT:    vperm v2, v3, v2, v4
+; P8-AIX-32-NEXT:    blr
 entry:
   %0 = load float, float* %f64, align 4
   %vecins = insertelement <4 x float> %vec, float %0, i32 0
@@ -285,6 +465,34 @@ define <2 x float> @s2v_test_f2(float* nocapture readonly %f64, <2 x float> %vec
 ; P8BE-NEXT:    lxsiwzx v3, 0, r3
 ; P8BE-NEXT:    vmrgow v2, v3, v2
 ; P8BE-NEXT:    blr
+;
+; P9-AIX-64-LABEL: s2v_test_f2:
+; P9-AIX-64:       # %bb.0: # %entry
+; P9-AIX-64-NEXT:    addi r3, r3, 4
+; P9-AIX-64-NEXT:    lxsiwzx v3, 0, r3
+; P9-AIX-64-NEXT:    vmrgow v2, v3, v2
+; P9-AIX-64-NEXT:    blr
+;
+; P9-AIX-32-LABEL: s2v_test_f2:
+; P9-AIX-32:       # %bb.0: # %entry
+; P9-AIX-32-NEXT:    lfs f0, 4(r3)
+; P9-AIX-32-NEXT:    xscvdpspn v3, f0
+; P9-AIX-32-NEXT:    vmrgow v2, v3, v2
+; P9-AIX-32-NEXT:    blr
+;
+; P8-AIX-64-LABEL: s2v_test_f2:
+; P8-AIX-64:       # %bb.0: # %entry
+; P8-AIX-64-NEXT:    addi r3, r3, 4
+; P8-AIX-64-NEXT:    lxsiwzx v3, 0, r3
+; P8-AIX-64-NEXT:    vmrgow v2, v3, v2
+; P8-AIX-64-NEXT:    blr
+;
+; P8-AIX-32-LABEL: s2v_test_f2:
+; P8-AIX-32:       # %bb.0: # %entry
+; P8-AIX-32-NEXT:    lfs f0, 4(r3)
+; P8-AIX-32-NEXT:    xscvdpspn v3, f0
+; P8-AIX-32-NEXT:    vmrgow v2, v3, v2
+; P8-AIX-32-NEXT:    blr
 entry:
   %arrayidx = getelementptr inbounds float, float* %f64, i64 1
   %0 = load float, float* %arrayidx, align 8
@@ -323,6 +531,36 @@ define <2 x float> @s2v_test_f3(float* nocapture readonly %f64, <2 x float> %vec
 ; P8BE-NEXT:    lxsiwzx v3, r3, r4
 ; P8BE-NEXT:    vmrgow v2, v3, v2
 ; P8BE-NEXT:    blr
+;
+; P9-AIX-64-LABEL: s2v_test_f3:
+; P9-AIX-64:       # %bb.0: # %entry
+; P9-AIX-64-NEXT:    sldi r4, r4, 2
+; P9-AIX-64-NEXT:    lxsiwzx v3, r3, r4
+; P9-AIX-64-NEXT:    vmrgow v2, v3, v2
+; P9-AIX-64-NEXT:    blr
+;
+; P9-AIX-32-LABEL: s2v_test_f3:
+; P9-AIX-32:       # %bb.0: # %entry
+; P9-AIX-32-NEXT:    slwi r4, r4, 2
+; P9-AIX-32-NEXT:    lfsx f0, r3, r4
+; P9-AIX-32-NEXT:    xscvdpspn v3, f0
+; P9-AIX-32-NEXT:    vmrgow v2, v3, v2
+; P9-AIX-32-NEXT:    blr
+;
+; P8-AIX-64-LABEL: s2v_test_f3:
+; P8-AIX-64:       # %bb.0: # %entry
+; P8-AIX-64-NEXT:    sldi r4, r4, 2
+; P8-AIX-64-NEXT:    lxsiwzx v3, r3, r4
+; P8-AIX-64-NEXT:    vmrgow v2, v3, v2
+; P8-AIX-64-NEXT:    blr
+;
+; P8-AIX-32-LABEL: s2v_test_f3:
+; P8-AIX-32:       # %bb.0: # %entry
+; P8-AIX-32-NEXT:    slwi r4, r4, 2
+; P8-AIX-32-NEXT:    lfsx f0, r3, r4
+; P8-AIX-32-NEXT:    xscvdpspn v3, f0
+; P8-AIX-32-NEXT:    vmrgow v2, v3, v2
+; P8-AIX-32-NEXT:    blr
 entry:
   %idxprom = sext i32 %Idx to i64
   %arrayidx = getelementptr inbounds float, float* %f64, i64 %idxprom
@@ -362,6 +600,34 @@ define <2 x float> @s2v_test_f4(float* nocapture readonly %f64, <2 x float> %vec
 ; P8BE-NEXT:    lxsiwzx v3, 0, r3
 ; P8BE-NEXT:    vmrgow v2, v3, v2
 ; P8BE-NEXT:    blr
+;
+; P9-AIX-64-LABEL: s2v_test_f4:
+; P9-AIX-64:       # %bb.0: # %entry
+; P9-AIX-64-NEXT:    addi r3, r3, 4
+; P9-AIX-64-NEXT:    lxsiwzx v3, 0, r3
+; P9-AIX-64-NEXT:    vmrgow v2, v3, v2
+; P9-AIX-64-NEXT:    blr
+;
+; P9-AIX-32-LABEL: s2v_test_f4:
+; P9-AIX-32:       # %bb.0: # %entry
+; P9-AIX-32-NEXT:    lfs f0, 4(r3)
+; P9-AIX-32-NEXT:    xscvdpspn v3, f0
+; P9-AIX-32-NEXT:    vmrgow v2, v3, v2
+; P9-AIX-32-NEXT:    blr
+;
+; P8-AIX-64-LABEL: s2v_test_f4:
+; P8-AIX-64:       # %bb.0: # %entry
+; P8-AIX-64-NEXT:    addi r3, r3, 4
+; P8-AIX-64-NEXT:    lxsiwzx v3, 0, r3
+; P8-AIX-64-NEXT:    vmrgow v2, v3, v2
+; P8-AIX-64-NEXT:    blr
+;
+; P8-AIX-32-LABEL: s2v_test_f4:
+; P8-AIX-32:       # %bb.0: # %entry
+; P8-AIX-32-NEXT:    lfs f0, 4(r3)
+; P8-AIX-32-NEXT:    xscvdpspn v3, f0
+; P8-AIX-32-NEXT:    vmrgow v2, v3, v2
+; P8-AIX-32-NEXT:    blr
 entry:
   %arrayidx = getelementptr inbounds float, float* %f64, i64 1
   %0 = load float, float* %arrayidx, align 8
@@ -396,6 +662,32 @@ define <2 x float> @s2v_test_f5(<2 x float> %vec, float* nocapture readonly %ptr
 ; P8BE-NEXT:    lxsiwzx v3, 0, r5
 ; P8BE-NEXT:    vmrgow v2, v3, v2
 ; P8BE-NEXT:    blr
+;
+; P9-AIX-64-LABEL: s2v_test_f5:
+; P9-AIX-64:       # %bb.0: # %entry
+; P9-AIX-64-NEXT:    lxsiwzx v3, 0, r3
+; P9-AIX-64-NEXT:    vmrgow v2, v3, v2
+; P9-AIX-64-NEXT:    blr
+;
+; P9-AIX-32-LABEL: s2v_test_f5:
+; P9-AIX-32:       # %bb.0: # %entry
+; P9-AIX-32-NEXT:    lfs f0, 0(r3)
+; P9-AIX-32-NEXT:    xscvdpspn v3, f0
+; P9-AIX-32-NEXT:    vmrgow v2, v3, v2
+; P9-AIX-32-NEXT:    blr
+;
+; P8-AIX-64-LABEL: s2v_test_f5:
+; P8-AIX-64:       # %bb.0: # %entry
+; P8-AIX-64-NEXT:    lxsiwzx v3, 0, r3
+; P8-AIX-64-NEXT:    vmrgow v2, v3, v2
+; P8-AIX-64-NEXT:    blr
+;
+; P8-AIX-32-LABEL: s2v_test_f5:
+; P8-AIX-32:       # %bb.0: # %entry
+; P8-AIX-32-NEXT:    lfs f0, 0(r3)
+; P8-AIX-32-NEXT:    xscvdpspn v3, f0
+; P8-AIX-32-NEXT:    vmrgow v2, v3, v2
+; P8-AIX-32-NEXT:    blr
 entry:
   %0 = load float, float* %ptr1, align 8
   %vecins = insertelement <2 x float> %vec, float %0, i32 0
diff --git a/llvm/test/CodeGen/PowerPC/vec_insert_elt.ll b/llvm/test/CodeGen/PowerPC/vec_insert_elt.ll
index 9bbb09ac3a96..225367f5a886 100644
--- a/llvm/test/CodeGen/PowerPC/vec_insert_elt.ll
+++ b/llvm/test/CodeGen/PowerPC/vec_insert_elt.ll
@@ -8,6 +8,12 @@
 ; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-linux-gnu \
 ; RUN:   -mcpu=pwr9 -ppc-asm-full-reg-names \
 ; RUN:   -ppc-vsr-nums-as-vr < %s | FileCheck %s --check-prefix=CHECK-P9
+; RUN: llc -mcpu=pwr8 -verify-machineinstrs -ppc-vsr-nums-as-vr \
+; RUN:   -ppc-asm-full-reg-names -mtriple=powerpc64-ibm-aix-xcoff < %s | \
+; RUN: FileCheck %s --check-prefixes=AIX-P8,AIX-P8-64
+; RUN: llc -mcpu=pwr8 -verify-machineinstrs -ppc-vsr-nums-as-vr \
+; RUN:   -ppc-asm-full-reg-names -mtriple=powerpc-ibm-aix-xcoff < %s | \
+; RUN: FileCheck %s --check-prefixes=AIX-P8,AIX-P8-32
 
 ; Byte indexed
 
@@ -30,6 +36,24 @@ define <16 x i8> @testByte(<16 x i8> %a, i64 %b, i64 %idx) {
 ; CHECK-P9-NEXT:    stbx r5, r4, r3
 ; CHECK-P9-NEXT:    lxv v2, -16(r1)
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-64-LABEL: testByte:
+; AIX-P8-64:       # %bb.0: # %entry
+; AIX-P8-64-NEXT:    addi r5, r1, -16
+; AIX-P8-64-NEXT:    clrldi r4, r4, 60
+; AIX-P8-64-NEXT:    stxvw4x v2, 0, r5
+; AIX-P8-64-NEXT:    stbx r3, r5, r4
+; AIX-P8-64-NEXT:    lxvw4x v2, 0, r5
+; AIX-P8-64-NEXT:    blr
+;
+; AIX-P8-32-LABEL: testByte:
+; AIX-P8-32:       # %bb.0: # %entry
+; AIX-P8-32-NEXT:    addi r3, r1, -16
+; AIX-P8-32-NEXT:    clrlwi r5, r6, 28
+; AIX-P8-32-NEXT:    stxvw4x v2, 0, r3
+; AIX-P8-32-NEXT:    stbx r4, r3, r5
+; AIX-P8-32-NEXT:    lxvw4x v2, 0, r3
+; AIX-P8-32-NEXT:    blr
 entry:
   %conv = trunc i64 %b to i8
   %vecins = insertelement <16 x i8> %a, i8 %conv, i64 %idx
@@ -59,6 +83,24 @@ define <8 x i16> @testHalf(<8 x i16> %a, i64 %b, i64 %idx) {
 ; CHECK-P9-NEXT:    sthx r5, r4, r3
 ; CHECK-P9-NEXT:    lxv v2, -16(r1)
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-64-LABEL: testHalf:
+; AIX-P8-64:       # %bb.0: # %entry
+; AIX-P8-64-NEXT:    addi r5, r1, -16
+; AIX-P8-64-NEXT:    rlwinm r4, r4, 1, 28, 30
+; AIX-P8-64-NEXT:    stxvw4x v2, 0, r5
+; AIX-P8-64-NEXT:    sthx r3, r5, r4
+; AIX-P8-64-NEXT:    lxvw4x v2, 0, r5
+; AIX-P8-64-NEXT:    blr
+;
+; AIX-P8-32-LABEL: testHalf:
+; AIX-P8-32:       # %bb.0: # %entry
+; AIX-P8-32-NEXT:    addi r3, r1, -16
+; AIX-P8-32-NEXT:    rlwinm r5, r6, 1, 28, 30
+; AIX-P8-32-NEXT:    stxvw4x v2, 0, r3
+; AIX-P8-32-NEXT:    sthx r4, r3, r5
+; AIX-P8-32-NEXT:    lxvw4x v2, 0, r3
+; AIX-P8-32-NEXT:    blr
 entry:
   %conv = trunc i64 %b to i16
   %vecins = insertelement <8 x i16> %a, i16 %conv, i64 %idx
@@ -88,6 +130,24 @@ define <4 x i32> @testWord(<4 x i32> %a, i64 %b, i64 %idx) {
 ; CHECK-P9-NEXT:    stwx r5, r4, r3
 ; CHECK-P9-NEXT:    lxv v2, -16(r1)
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-64-LABEL: testWord:
+; AIX-P8-64:       # %bb.0: # %entry
+; AIX-P8-64-NEXT:    addi r5, r1, -16
+; AIX-P8-64-NEXT:    rlwinm r4, r4, 2, 28, 29
+; AIX-P8-64-NEXT:    stxvw4x v2, 0, r5
+; AIX-P8-64-NEXT:    stwx r3, r5, r4
+; AIX-P8-64-NEXT:    lxvw4x v2, 0, r5
+; AIX-P8-64-NEXT:    blr
+;
+; AIX-P8-32-LABEL: testWord:
+; AIX-P8-32:       # %bb.0: # %entry
+; AIX-P8-32-NEXT:    addi r3, r1, -16
+; AIX-P8-32-NEXT:    rlwinm r5, r6, 2, 28, 29
+; AIX-P8-32-NEXT:    stxvw4x v2, 0, r3
+; AIX-P8-32-NEXT:    stwx r4, r3, r5
+; AIX-P8-32-NEXT:    lxvw4x v2, 0, r3
+; AIX-P8-32-NEXT:    blr
 entry:
   %conv = trunc i64 %b to i32
   %vecins = insertelement <4 x i32> %a, i32 %conv, i64 %idx
@@ -115,6 +175,30 @@ define <4 x i32> @testWordImm(<4 x i32> %a, i64 %b) {
 ; CHECK-P9-NEXT:    xxinsertw v2, vs0, 4
 ; CHECK-P9-NEXT:    xxinsertw v2, vs0, 12
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-64-LABEL: testWordImm:
+; AIX-P8-64:       # %bb.0: # %entry
+; AIX-P8-64-NEXT:    ld r4, L..C0(r2) # %const.0
+; AIX-P8-64-NEXT:    mtvsrwz v4, r3
+; AIX-P8-64-NEXT:    ld r3, L..C1(r2) # %const.1
+; AIX-P8-64-NEXT:    lxvw4x v3, 0, r4
+; AIX-P8-64-NEXT:    vperm v2, v2, v4, v3
+; AIX-P8-64-NEXT:    lxvw4x v3, 0, r3
+; AIX-P8-64-NEXT:    vperm v2, v2, v4, v3
+; AIX-P8-64-NEXT:    blr
+;
+; AIX-P8-32-LABEL: testWordImm:
+; AIX-P8-32:       # %bb.0: # %entry
+; AIX-P8-32-NEXT:    lwz r3, L..C0(r2) # %const.0
+; AIX-P8-32-NEXT:    stw r4, -16(r1)
+; AIX-P8-32-NEXT:    addi r4, r1, -16
+; AIX-P8-32-NEXT:    lxvw4x v4, 0, r4
+; AIX-P8-32-NEXT:    lxvw4x v3, 0, r3
+; AIX-P8-32-NEXT:    lwz r3, L..C1(r2) # %const.1
+; AIX-P8-32-NEXT:    vperm v2, v2, v4, v3
+; AIX-P8-32-NEXT:    lxvw4x v3, 0, r3
+; AIX-P8-32-NEXT:    vperm v2, v2, v4, v3
+; AIX-P8-32-NEXT:    blr
 entry:
   %conv = trunc i64 %b to i32
   %vecins = insertelement <4 x i32> %a, i32 %conv, i32 1
@@ -145,6 +229,31 @@ define <2 x i64> @testDoubleword(<2 x i64> %a, i64 %b, i64 %idx) {
 ; CHECK-P9-NEXT:    stdx r5, r4, r3
 ; CHECK-P9-NEXT:    lxv v2, -16(r1)
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-64-LABEL: testDoubleword:
+; AIX-P8-64:       # %bb.0: # %entry
+; AIX-P8-64-NEXT:    addi r5, r1, -16
+; AIX-P8-64-NEXT:    rlwinm r4, r4, 3, 28, 28
+; AIX-P8-64-NEXT:    stxvd2x v2, 0, r5
+; AIX-P8-64-NEXT:    stdx r3, r5, r4
+; AIX-P8-64-NEXT:    lxvd2x v2, 0, r5
+; AIX-P8-64-NEXT:    blr
+;
+; AIX-P8-32-LABEL: testDoubleword:
+; AIX-P8-32:       # %bb.0: # %entry
+; AIX-P8-32-NEXT:    add r6, r6, r6
+; AIX-P8-32-NEXT:    addi r5, r1, -32
+; AIX-P8-32-NEXT:    rlwinm r7, r6, 2, 28, 29
+; AIX-P8-32-NEXT:    stxvw4x v2, 0, r5
+; AIX-P8-32-NEXT:    stwx r3, r5, r7
+; AIX-P8-32-NEXT:    addi r3, r6, 1
+; AIX-P8-32-NEXT:    lxvw4x vs0, 0, r5
+; AIX-P8-32-NEXT:    addi r5, r1, -16
+; AIX-P8-32-NEXT:    rlwinm r3, r3, 2, 28, 29
+; AIX-P8-32-NEXT:    stxvw4x vs0, 0, r5
+; AIX-P8-32-NEXT:    stwx r4, r5, r3
+; AIX-P8-32-NEXT:    lxvw4x v2, 0, r5
+; AIX-P8-32-NEXT:    blr
 entry:
   %vecins = insertelement <2 x i64> %a, i64 %b, i64 %idx
   ret <2 x i64> %vecins
@@ -168,6 +277,28 @@ define <2 x i64> @testDoublewordImm(<2 x i64> %a, i64 %b) {
 ; CHECK-P9-NEXT:    mtfprd f0, r5
 ; CHECK-P9-NEXT:    xxmrghd v2, v2, vs0
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-64-LABEL: testDoublewordImm:
+; AIX-P8-64:       # %bb.0: # %entry
+; AIX-P8-64-NEXT:    mtfprd f0, r3
+; AIX-P8-64-NEXT:    xxmrghd v2, v2, vs0
+; AIX-P8-64-NEXT:    blr
+;
+; AIX-P8-32-LABEL: testDoublewordImm:
+; AIX-P8-32:       # %bb.0: # %entry
+; AIX-P8-32-NEXT:    lwz r5, L..C2(r2) # %const.0
+; AIX-P8-32-NEXT:    stw r3, -16(r1)
+; AIX-P8-32-NEXT:    stw r4, -32(r1)
+; AIX-P8-32-NEXT:    addi r3, r1, -16
+; AIX-P8-32-NEXT:    addi r4, r1, -32
+; AIX-P8-32-NEXT:    lxvw4x v4, 0, r3
+; AIX-P8-32-NEXT:    lwz r3, L..C3(r2) # %const.1
+; AIX-P8-32-NEXT:    lxvw4x v3, 0, r5
+; AIX-P8-32-NEXT:    vperm v2, v2, v4, v3
+; AIX-P8-32-NEXT:    lxvw4x v3, 0, r3
+; AIX-P8-32-NEXT:    lxvw4x v4, 0, r4
+; AIX-P8-32-NEXT:    vperm v2, v2, v4, v3
+; AIX-P8-32-NEXT:    blr
 entry:
   %vecins = insertelement <2 x i64> %a, i64 %b, i32 1
   ret <2 x i64> %vecins
@@ -189,6 +320,28 @@ define <2 x i64> @testDoublewordImm2(<2 x i64> %a, i64 %b) {
 ; CHECK-P9-NEXT:    mtfprd f0, r5
 ; CHECK-P9-NEXT:    xxpermdi v2, vs0, v2, 1
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-64-LABEL: testDoublewordImm2:
+; AIX-P8-64:       # %bb.0: # %entry
+; AIX-P8-64-NEXT:    mtfprd f0, r3
+; AIX-P8-64-NEXT:    xxpermdi v2, vs0, v2, 1
+; AIX-P8-64-NEXT:    blr
+;
+; AIX-P8-32-LABEL: testDoublewordImm2:
+; AIX-P8-32:       # %bb.0: # %entry
+; AIX-P8-32-NEXT:    lwz r5, L..C4(r2) # %const.0
+; AIX-P8-32-NEXT:    stw r3, -16(r1)
+; AIX-P8-32-NEXT:    stw r4, -32(r1)
+; AIX-P8-32-NEXT:    addi r3, r1, -16
+; AIX-P8-32-NEXT:    addi r4, r1, -32
+; AIX-P8-32-NEXT:    lxvw4x v4, 0, r3
+; AIX-P8-32-NEXT:    lwz r3, L..C5(r2) # %const.1
+; AIX-P8-32-NEXT:    lxvw4x v3, 0, r5
+; AIX-P8-32-NEXT:    vperm v2, v4, v2, v3
+; AIX-P8-32-NEXT:    lxvw4x v3, 0, r3
+; AIX-P8-32-NEXT:    lxvw4x v4, 0, r4
+; AIX-P8-32-NEXT:    vperm v2, v2, v4, v3
+; AIX-P8-32-NEXT:    blr
 entry:
   %vecins = insertelement <2 x i64> %a, i64 %b, i32 0
   ret <2 x i64> %vecins
@@ -221,6 +374,15 @@ define <4 x float> @testFloat1(<4 x float> %a, float %b, i32 zeroext %idx1) {
 ; CHECK-P9-NEXT:    stfsx f1, r4, r3
 ; CHECK-P9-NEXT:    lxv v2, -16(r1)
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-LABEL: testFloat1:
+; AIX-P8:       # %bb.0: # %entry
+; AIX-P8-NEXT:    addi r3, r1, -16
+; AIX-P8-NEXT:    rlwinm r4, r4, 2, 28, 29
+; AIX-P8-NEXT:    stxvw4x v2, 0, r3
+; AIX-P8-NEXT:    stfsx f1, r3, r4
+; AIX-P8-NEXT:    lxvw4x v2, 0, r3
+; AIX-P8-NEXT:    blr
 entry:
   %vecins = insertelement <4 x float> %a, float %b, i32 %idx1
   ret <4 x float> %vecins
@@ -266,6 +428,38 @@ define <4 x float> @testFloat2(<4 x float> %a, i8* %b, i32 zeroext %idx1, i32 ze
 ; CHECK-P9-NEXT:    stwx r3, r5, r4
 ; CHECK-P9-NEXT:    lxv v2, -32(r1)
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-64-LABEL: testFloat2:
+; AIX-P8-64:       # %bb.0: # %entry
+; AIX-P8-64-NEXT:    lwz r7, 0(r3)
+; AIX-P8-64-NEXT:    addi r6, r1, -32
+; AIX-P8-64-NEXT:    rlwinm r4, r4, 2, 28, 29
+; AIX-P8-64-NEXT:    rlwinm r5, r5, 2, 28, 29
+; AIX-P8-64-NEXT:    stxvw4x v2, 0, r6
+; AIX-P8-64-NEXT:    stwx r7, r6, r4
+; AIX-P8-64-NEXT:    addi r4, r1, -16
+; AIX-P8-64-NEXT:    lxvw4x vs0, 0, r6
+; AIX-P8-64-NEXT:    lwz r3, 1(r3)
+; AIX-P8-64-NEXT:    stxvw4x vs0, 0, r4
+; AIX-P8-64-NEXT:    stwx r3, r4, r5
+; AIX-P8-64-NEXT:    lxvw4x v2, 0, r4
+; AIX-P8-64-NEXT:    blr
+;
+; AIX-P8-32-LABEL: testFloat2:
+; AIX-P8-32:       # %bb.0: # %entry
+; AIX-P8-32-NEXT:    lwz r7, 0(r3)
+; AIX-P8-32-NEXT:    addi r6, r1, -32
+; AIX-P8-32-NEXT:    rlwinm r4, r4, 2, 28, 29
+; AIX-P8-32-NEXT:    stxvw4x v2, 0, r6
+; AIX-P8-32-NEXT:    stwx r7, r6, r4
+; AIX-P8-32-NEXT:    rlwinm r4, r5, 2, 28, 29
+; AIX-P8-32-NEXT:    addi r5, r1, -16
+; AIX-P8-32-NEXT:    lxvw4x vs0, 0, r6
+; AIX-P8-32-NEXT:    lwz r3, 1(r3)
+; AIX-P8-32-NEXT:    stxvw4x vs0, 0, r5
+; AIX-P8-32-NEXT:    stwx r3, r5, r4
+; AIX-P8-32-NEXT:    lxvw4x v2, 0, r5
+; AIX-P8-32-NEXT:    blr
 entry:
   %0 = bitcast i8* %b to float*
   %add.ptr1 = getelementptr inbounds i8, i8* %b, i64 1
@@ -324,6 +518,42 @@ define <4 x float> @testFloat3(<4 x float> %a, i8* %b, i32 zeroext %idx1, i32 ze
 ; CHECK-P9-NEXT:    stwx r3, r5, r4
 ; CHECK-P9-NEXT:    lxv v2, -32(r1)
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-64-LABEL: testFloat3:
+; AIX-P8-64:       # %bb.0: # %entry
+; AIX-P8-64-NEXT:    lis r6, 1
+; AIX-P8-64-NEXT:    addi r7, r1, -32
+; AIX-P8-64-NEXT:    rlwinm r4, r4, 2, 28, 29
+; AIX-P8-64-NEXT:    rlwinm r5, r5, 2, 28, 29
+; AIX-P8-64-NEXT:    lwzx r6, r3, r6
+; AIX-P8-64-NEXT:    stxvw4x v2, 0, r7
+; AIX-P8-64-NEXT:    stwx r6, r7, r4
+; AIX-P8-64-NEXT:    li r4, 1
+; AIX-P8-64-NEXT:    lxvw4x vs0, 0, r7
+; AIX-P8-64-NEXT:    rldic r4, r4, 36, 27
+; AIX-P8-64-NEXT:    lwzx r3, r3, r4
+; AIX-P8-64-NEXT:    addi r4, r1, -16
+; AIX-P8-64-NEXT:    stxvw4x vs0, 0, r4
+; AIX-P8-64-NEXT:    stwx r3, r4, r5
+; AIX-P8-64-NEXT:    lxvw4x v2, 0, r4
+; AIX-P8-64-NEXT:    blr
+;
+; AIX-P8-32-LABEL: testFloat3:
+; AIX-P8-32:       # %bb.0: # %entry
+; AIX-P8-32-NEXT:    lis r6, 1
+; AIX-P8-32-NEXT:    rlwinm r4, r4, 2, 28, 29
+; AIX-P8-32-NEXT:    addi r7, r1, -32
+; AIX-P8-32-NEXT:    lwzx r6, r3, r6
+; AIX-P8-32-NEXT:    stxvw4x v2, 0, r7
+; AIX-P8-32-NEXT:    stwx r6, r7, r4
+; AIX-P8-32-NEXT:    rlwinm r4, r5, 2, 28, 29
+; AIX-P8-32-NEXT:    addi r5, r1, -16
+; AIX-P8-32-NEXT:    lxvw4x vs0, 0, r7
+; AIX-P8-32-NEXT:    lwz r3, 0(r3)
+; AIX-P8-32-NEXT:    stxvw4x vs0, 0, r5
+; AIX-P8-32-NEXT:    stwx r3, r5, r4
+; AIX-P8-32-NEXT:    lxvw4x v2, 0, r5
+; AIX-P8-32-NEXT:    blr
 entry:
   %add.ptr = getelementptr inbounds i8, i8* %b, i64 65536
   %0 = bitcast i8* %add.ptr to float*
@@ -359,6 +589,28 @@ define <4 x float> @testFloatImm1(<4 x float> %a, float %b) {
 ; CHECK-P9-NEXT:    xxinsertw v2, vs0, 0
 ; CHECK-P9-NEXT:    xxinsertw v2, vs0, 8
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-64-LABEL: testFloatImm1:
+; AIX-P8-64:       # %bb.0: # %entry
+; AIX-P8-64-NEXT:    ld r3, L..C2(r2) # %const.0
+; AIX-P8-64-NEXT:    xscvdpspn v3, f1
+; AIX-P8-64-NEXT:    lxvw4x v4, 0, r3
+; AIX-P8-64-NEXT:    ld r3, L..C3(r2) # %const.1
+; AIX-P8-64-NEXT:    vperm v2, v3, v2, v4
+; AIX-P8-64-NEXT:    lxvw4x v4, 0, r3
+; AIX-P8-64-NEXT:    vperm v2, v2, v3, v4
+; AIX-P8-64-NEXT:    blr
+;
+; AIX-P8-32-LABEL: testFloatImm1:
+; AIX-P8-32:       # %bb.0: # %entry
+; AIX-P8-32-NEXT:    lwz r3, L..C6(r2) # %const.0
+; AIX-P8-32-NEXT:    xscvdpspn v3, f1
+; AIX-P8-32-NEXT:    lxvw4x v4, 0, r3
+; AIX-P8-32-NEXT:    lwz r3, L..C7(r2) # %const.1
+; AIX-P8-32-NEXT:    vperm v2, v3, v2, v4
+; AIX-P8-32-NEXT:    lxvw4x v4, 0, r3
+; AIX-P8-32-NEXT:    vperm v2, v2, v3, v4
+; AIX-P8-32-NEXT:    blr
 entry:
   %vecins = insertelement <4 x float> %a, float %b, i32 0
   %vecins1 = insertelement <4 x float> %vecins, float %b, i32 2
@@ -391,6 +643,33 @@ define <4 x float> @testFloatImm2(<4 x float> %a, i32* %b) {
 ; CHECK-P9-NEXT:    mtfprwz f0, r3
 ; CHECK-P9-NEXT:    xxinsertw v2, vs0, 8
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-64-LABEL: testFloatImm2:
+; AIX-P8-64:       # %bb.0: # %entry
+; AIX-P8-64-NEXT:    ld r4, L..C4(r2) # %const.0
+; AIX-P8-64-NEXT:    lxsiwzx v3, 0, r3
+; AIX-P8-64-NEXT:    li r5, 4
+; AIX-P8-64-NEXT:    lxvw4x v4, 0, r4
+; AIX-P8-64-NEXT:    ld r4, L..C5(r2) # %const.1
+; AIX-P8-64-NEXT:    vperm v2, v3, v2, v4
+; AIX-P8-64-NEXT:    lxsiwzx v3, r3, r5
+; AIX-P8-64-NEXT:    lxvw4x v4, 0, r4
+; AIX-P8-64-NEXT:    vperm v2, v2, v3, v4
+; AIX-P8-64-NEXT:    blr
+;
+; AIX-P8-32-LABEL: testFloatImm2:
+; AIX-P8-32:       # %bb.0: # %entry
+; AIX-P8-32-NEXT:    lfs f0, 0(r3)
+; AIX-P8-32-NEXT:    lwz r4, L..C8(r2) # %const.0
+; AIX-P8-32-NEXT:    xscvdpspn v3, f0
+; AIX-P8-32-NEXT:    lxvw4x v4, 0, r4
+; AIX-P8-32-NEXT:    lfs f0, 4(r3)
+; AIX-P8-32-NEXT:    lwz r3, L..C9(r2) # %const.1
+; AIX-P8-32-NEXT:    vperm v2, v3, v2, v4
+; AIX-P8-32-NEXT:    lxvw4x v4, 0, r3
+; AIX-P8-32-NEXT:    xscvdpspn v3, f0
+; AIX-P8-32-NEXT:    vperm v2, v2, v3, v4
+; AIX-P8-32-NEXT:    blr
 entry:
   %0 = bitcast i32* %b to float*
   %add.ptr1 = getelementptr inbounds i32, i32* %b, i64 1
@@ -435,6 +714,36 @@ define <4 x float> @testFloatImm3(<4 x float> %a, i32* %b) {
 ; CHECK-P9-NEXT:    mtfprwz f0, r3
 ; CHECK-P9-NEXT:    xxinsertw v2, vs0, 8
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-64-LABEL: testFloatImm3:
+; AIX-P8-64:       # %bb.0: # %entry
+; AIX-P8-64-NEXT:    ld r4, L..C6(r2) # %const.0
+; AIX-P8-64-NEXT:    lis r5, 4
+; AIX-P8-64-NEXT:    lxsiwzx v3, r3, r5
+; AIX-P8-64-NEXT:    li r5, 1
+; AIX-P8-64-NEXT:    rldic r5, r5, 38, 25
+; AIX-P8-64-NEXT:    lxvw4x v4, 0, r4
+; AIX-P8-64-NEXT:    ld r4, L..C7(r2) # %const.1
+; AIX-P8-64-NEXT:    vperm v2, v3, v2, v4
+; AIX-P8-64-NEXT:    lxsiwzx v3, r3, r5
+; AIX-P8-64-NEXT:    lxvw4x v4, 0, r4
+; AIX-P8-64-NEXT:    vperm v2, v2, v3, v4
+; AIX-P8-64-NEXT:    blr
+;
+; AIX-P8-32-LABEL: testFloatImm3:
+; AIX-P8-32:       # %bb.0: # %entry
+; AIX-P8-32-NEXT:    lis r4, 4
+; AIX-P8-32-NEXT:    lfsx f0, r3, r4
+; AIX-P8-32-NEXT:    lwz r4, L..C10(r2) # %const.0
+; AIX-P8-32-NEXT:    xscvdpspn v3, f0
+; AIX-P8-32-NEXT:    lxvw4x v4, 0, r4
+; AIX-P8-32-NEXT:    lfs f0, 0(r3)
+; AIX-P8-32-NEXT:    lwz r3, L..C11(r2) # %const.1
+; AIX-P8-32-NEXT:    vperm v2, v3, v2, v4
+; AIX-P8-32-NEXT:    lxvw4x v4, 0, r3
+; AIX-P8-32-NEXT:    xscvdpspn v3, f0
+; AIX-P8-32-NEXT:    vperm v2, v2, v3, v4
+; AIX-P8-32-NEXT:    blr
 entry:
   %add.ptr = getelementptr inbounds i32, i32* %b, i64 65536
   %0 = bitcast i32* %add.ptr to float*
@@ -474,6 +783,24 @@ define <2 x double> @testDouble1(<2 x double> %a, double %b, i32 zeroext %idx1)
 ; CHECK-P9-NEXT:    stfdx f1, r4, r3
 ; CHECK-P9-NEXT:    lxv v2, -16(r1)
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-64-LABEL: testDouble1:
+; AIX-P8-64:       # %bb.0: # %entry
+; AIX-P8-64-NEXT:    addi r3, r1, -16
+; AIX-P8-64-NEXT:    rlwinm r4, r4, 3, 28, 28
+; AIX-P8-64-NEXT:    stxvd2x v2, 0, r3
+; AIX-P8-64-NEXT:    stfdx f1, r3, r4
+; AIX-P8-64-NEXT:    lxvd2x v2, 0, r3
+; AIX-P8-64-NEXT:    blr
+;
+; AIX-P8-32-LABEL: testDouble1:
+; AIX-P8-32:       # %bb.0: # %entry
+; AIX-P8-32-NEXT:    addi r3, r1, -16
+; AIX-P8-32-NEXT:    rlwinm r4, r5, 3, 28, 28
+; AIX-P8-32-NEXT:    stxvd2x v2, 0, r3
+; AIX-P8-32-NEXT:    stfdx f1, r3, r4
+; AIX-P8-32-NEXT:    lxvd2x v2, 0, r3
+; AIX-P8-32-NEXT:    blr
 entry:
   %vecins = insertelement <2 x double> %a, double %b, i32 %idx1
   ret <2 x double> %vecins
@@ -520,6 +847,39 @@ define <2 x double> @testDouble2(<2 x double> %a, i8* %b, i32 zeroext %idx1, i32
 ; CHECK-P9-NEXT:    stdx r3, r5, r4
 ; CHECK-P9-NEXT:    lxv v2, -16(r1)
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-64-LABEL: testDouble2:
+; AIX-P8-64:       # %bb.0: # %entry
+; AIX-P8-64-NEXT:    ld r7, 0(r3)
+; AIX-P8-64-NEXT:    addi r6, r1, -32
+; AIX-P8-64-NEXT:    rlwinm r4, r4, 3, 28, 28
+; AIX-P8-64-NEXT:    rlwinm r5, r5, 3, 28, 28
+; AIX-P8-64-NEXT:    stxvd2x v2, 0, r6
+; AIX-P8-64-NEXT:    stdx r7, r6, r4
+; AIX-P8-64-NEXT:    li r4, 1
+; AIX-P8-64-NEXT:    lxvd2x vs0, 0, r6
+; AIX-P8-64-NEXT:    ldx r3, r3, r4
+; AIX-P8-64-NEXT:    addi r4, r1, -16
+; AIX-P8-64-NEXT:    stxvd2x vs0, 0, r4
+; AIX-P8-64-NEXT:    stdx r3, r4, r5
+; AIX-P8-64-NEXT:    lxvd2x v2, 0, r4
+; AIX-P8-64-NEXT:    blr
+;
+; AIX-P8-32-LABEL: testDouble2:
+; AIX-P8-32:       # %bb.0: # %entry
+; AIX-P8-32-NEXT:    lfd f0, 0(r3)
+; AIX-P8-32-NEXT:    addi r6, r1, -32
+; AIX-P8-32-NEXT:    rlwinm r4, r4, 3, 28, 28
+; AIX-P8-32-NEXT:    stxvd2x v2, 0, r6
+; AIX-P8-32-NEXT:    stfdx f0, r6, r4
+; AIX-P8-32-NEXT:    addi r4, r1, -16
+; AIX-P8-32-NEXT:    lxvd2x vs0, 0, r6
+; AIX-P8-32-NEXT:    lfd f1, 1(r3)
+; AIX-P8-32-NEXT:    rlwinm r3, r5, 3, 28, 28
+; AIX-P8-32-NEXT:    stxvd2x vs0, 0, r4
+; AIX-P8-32-NEXT:    stfdx f1, r4, r3
+; AIX-P8-32-NEXT:    lxvd2x v2, 0, r4
+; AIX-P8-32-NEXT:    blr
 entry:
   %0 = bitcast i8* %b to double*
   %add.ptr1 = getelementptr inbounds i8, i8* %b, i64 1
@@ -578,6 +938,42 @@ define <2 x double> @testDouble3(<2 x double> %a, i8* %b, i32 zeroext %idx1, i32
 ; CHECK-P9-NEXT:    stdx r3, r5, r4
 ; CHECK-P9-NEXT:    lxv v2, -16(r1)
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-64-LABEL: testDouble3:
+; AIX-P8-64:       # %bb.0: # %entry
+; AIX-P8-64-NEXT:    lis r6, 1
+; AIX-P8-64-NEXT:    addi r7, r1, -32
+; AIX-P8-64-NEXT:    rlwinm r4, r4, 3, 28, 28
+; AIX-P8-64-NEXT:    li r8, 1
+; AIX-P8-64-NEXT:    rlwinm r5, r5, 3, 28, 28
+; AIX-P8-64-NEXT:    ldx r6, r3, r6
+; AIX-P8-64-NEXT:    stxvd2x v2, 0, r7
+; AIX-P8-64-NEXT:    stdx r6, r7, r4
+; AIX-P8-64-NEXT:    rldic r4, r8, 36, 27
+; AIX-P8-64-NEXT:    lxvd2x vs0, 0, r7
+; AIX-P8-64-NEXT:    ldx r3, r3, r4
+; AIX-P8-64-NEXT:    addi r4, r1, -16
+; AIX-P8-64-NEXT:    stxvd2x vs0, 0, r4
+; AIX-P8-64-NEXT:    stdx r3, r4, r5
+; AIX-P8-64-NEXT:    lxvd2x v2, 0, r4
+; AIX-P8-64-NEXT:    blr
+;
+; AIX-P8-32-LABEL: testDouble3:
+; AIX-P8-32:       # %bb.0: # %entry
+; AIX-P8-32-NEXT:    lis r6, 1
+; AIX-P8-32-NEXT:    rlwinm r4, r4, 3, 28, 28
+; AIX-P8-32-NEXT:    lfdx f0, r3, r6
+; AIX-P8-32-NEXT:    addi r6, r1, -32
+; AIX-P8-32-NEXT:    stxvd2x v2, 0, r6
+; AIX-P8-32-NEXT:    stfdx f0, r6, r4
+; AIX-P8-32-NEXT:    addi r4, r1, -16
+; AIX-P8-32-NEXT:    lxvd2x vs0, 0, r6
+; AIX-P8-32-NEXT:    lfd f1, 0(r3)
+; AIX-P8-32-NEXT:    rlwinm r3, r5, 3, 28, 28
+; AIX-P8-32-NEXT:    stxvd2x vs0, 0, r4
+; AIX-P8-32-NEXT:    stfdx f1, r4, r3
+; AIX-P8-32-NEXT:    lxvd2x v2, 0, r4
+; AIX-P8-32-NEXT:    blr
 entry:
   %add.ptr = getelementptr inbounds i8, i8* %b, i64 65536
   %0 = bitcast i8* %add.ptr to double*
@@ -610,6 +1006,12 @@ define <2 x double> @testDoubleImm1(<2 x double> %a, double %b) {
 ; CHECK-P9-NEXT:    # kill: def $f1 killed $f1 def $vsl1
 ; CHECK-P9-NEXT:    xxpermdi v2, vs1, v2, 1
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-LABEL: testDoubleImm1:
+; AIX-P8:       # %bb.0: # %entry
+; AIX-P8-NEXT:    # kill: def $f1 killed $f1 def $vsl1
+; AIX-P8-NEXT:    xxpermdi v2, vs1, v2, 1
+; AIX-P8-NEXT:    blr
 entry:
   %vecins = insertelement <2 x double> %a, double %b, i32 0
   ret <2 x double> %vecins
@@ -633,6 +1035,12 @@ define <2 x double> @testDoubleImm2(<2 x double> %a, i32* %b) {
 ; CHECK-P9-NEXT:    lfd f0, 0(r5)
 ; CHECK-P9-NEXT:    xxpermdi v2, vs0, v2, 1
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-LABEL: testDoubleImm2:
+; AIX-P8:       # %bb.0: # %entry
+; AIX-P8-NEXT:    lfd f0, 0(r3)
+; AIX-P8-NEXT:    xxpermdi v2, vs0, v2, 1
+; AIX-P8-NEXT:    blr
 entry:
   %0 = bitcast i32* %b to double*
   %1 = load double, double* %0, align 8
@@ -658,6 +1066,12 @@ define <2 x double> @testDoubleImm3(<2 x double> %a, i32* %b) {
 ; CHECK-P9-NEXT:    lfd f0, 4(r5)
 ; CHECK-P9-NEXT:    xxpermdi v2, vs0, v2, 1
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-LABEL: testDoubleImm3:
+; AIX-P8:       # %bb.0: # %entry
+; AIX-P8-NEXT:    lfd f0, 4(r3)
+; AIX-P8-NEXT:    xxpermdi v2, vs0, v2, 1
+; AIX-P8-NEXT:    blr
 entry:
   %add.ptr = getelementptr inbounds i32, i32* %b, i64 1
   %0 = bitcast i32* %add.ptr to double*
@@ -685,6 +1099,13 @@ define <2 x double> @testDoubleImm4(<2 x double> %a, i32* %b) {
 ; CHECK-P9-NEXT:    lfdx f0, r5, r3
 ; CHECK-P9-NEXT:    xxpermdi v2, vs0, v2, 1
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-LABEL: testDoubleImm4:
+; AIX-P8:       # %bb.0: # %entry
+; AIX-P8-NEXT:    lis r4, 4
+; AIX-P8-NEXT:    lfdx f0, r3, r4
+; AIX-P8-NEXT:    xxpermdi v2, vs0, v2, 1
+; AIX-P8-NEXT:    blr
 entry:
   %add.ptr = getelementptr inbounds i32, i32* %b, i64 65536
   %0 = bitcast i32* %add.ptr to double*
@@ -717,6 +1138,20 @@ define <2 x double> @testDoubleImm5(<2 x double> %a, i32* %b) {
 ; CHECK-P9-NEXT:    lfdx f0, r5, r3
 ; CHECK-P9-NEXT:    xxpermdi v2, vs0, v2, 1
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-64-LABEL: testDoubleImm5:
+; AIX-P8-64:       # %bb.0: # %entry
+; AIX-P8-64-NEXT:    li r4, 1
+; AIX-P8-64-NEXT:    rldic r4, r4, 38, 25
+; AIX-P8-64-NEXT:    lfdx f0, r3, r4
+; AIX-P8-64-NEXT:    xxpermdi v2, vs0, v2, 1
+; AIX-P8-64-NEXT:    blr
+;
+; AIX-P8-32-LABEL: testDoubleImm5:
+; AIX-P8-32:       # %bb.0: # %entry
+; AIX-P8-32-NEXT:    lfd f0, 0(r3)
+; AIX-P8-32-NEXT:    xxpermdi v2, vs0, v2, 1
+; AIX-P8-32-NEXT:    blr
 entry:
   %add.ptr = getelementptr inbounds i32, i32* %b, i64 68719476736
   %0 = bitcast i32* %add.ptr to double*
@@ -743,6 +1178,24 @@ define dso_local <4 x float> @testInsertDoubleToFloat(<4 x float> %a, double %b)
 ; CHECK-P9-NEXT:    xscvdpsp f0, f1
 ; CHECK-P9-NEXT:    xxinsertw v2, vs0, 4
 ; CHECK-P9-NEXT:    blr
+;
+; AIX-P8-64-LABEL: testInsertDoubleToFloat:
+; AIX-P8-64:       # %bb.0: # %entry
+; AIX-P8-64-NEXT:    xsrsp f0, f1
+; AIX-P8-64-NEXT:    ld r3, L..C8(r2) # %const.0
+; AIX-P8-64-NEXT:    lxvw4x v4, 0, r3
+; AIX-P8-64-NEXT:    xscvdpspn v3, f0
+; AIX-P8-64-NEXT:    vperm v2, v2, v3, v4
+; AIX-P8-64-NEXT:    blr
+;
+; AIX-P8-32-LABEL: testInsertDoubleToFloat:
+; AIX-P8-32:       # %bb.0: # %entry
+; AIX-P8-32-NEXT:    xsrsp f0, f1
+; AIX-P8-32-NEXT:    lwz r3, L..C12(r2) # %const.0
+; AIX-P8-32-NEXT:    lxvw4x v4, 0, r3
+; AIX-P8-32-NEXT:    xscvdpspn v3, f0
+; AIX-P8-32-NEXT:    vperm v2, v2, v3, v4
+; AIX-P8-32-NEXT:    blr
 entry:
   %conv = fptrunc double %b to float
   %vecins = insertelement <4 x float> %a, float %conv, i32 1