[PowerPC][Power10] Exploit store rightmost vector element instructions.

[Conanap]

Summary

Using the store rightmost vector element instructions to do vector
element extraction and store. The rightmost vector element on little
endian is the zeroth vector element, with these patterns that element
can be extracted and stored in one instruction for all vector types.

[lei]

llvm/test/CodeGen/PowerPC/builtins-ppc-p10vsx.ll

1

There's some redundancies here:

1. The default CHECK are not used in any of the run lines, but are updated in some cases.
2. The newly generated CHECK-LE and CHECK-BE are at times exact duplicate of each other and of the default CHECK.

Suggest to update thus so it's more clear what the changes are:

RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu \
; RUN:   -mcpu=pwr10 -ppc-asm-full-reg-names -ppc-vsr-nums-as-vr < %s | \
; RUN:   FileCheck %s --check-prefix=CHECK,LE
; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-linux-gnu \
; RUN:   -mcpu=pwr10 -ppc-asm-full-reg-names -ppc-vsr-nums-as-vr < %s | \
; RUN:   FileCheck %s --check-prefix=CHECK,BE

For cases where there is a differnce in LE and BE behaviour, remove the default CHECK and add new. Otherwise just leave the original default CHECKs.

llvm/test/CodeGen/PowerPC/store-rightmost-vector-elt.ll

4

run line for BE?

[amyk]

Please address the comments regarding the test cases.

llvm/test/CodeGen/PowerPC/builtins-ppc-p10vsx.ll

1

I also agree with Lei's comment. We should add check prefixes and remove/add the CHECKs when its necessary.

llvm/test/CodeGen/PowerPC/store-rightmost-vector-elt.ll

4

Agree we should have a run line for BE.

[Conanap]

Will take over patch from Kamau

[Conanap]

Updated tests as per Amy and Lei's comments

[lei]

LGTM

[amyk]

LGTM. Thanks for the update Albion.

[jsji]

In D89195#2461459, @amyk wrote:

LGTM. Thanks for the update Albion.

@amyk If you request changes on behalf of PowerPC, you should accept it on behalf of PowerPC as well.
Or else it will show This revision was not accepted when it landed; it landed in state Needs Review. Thanks.

[nemanjai]

llvm/lib/Target/PowerPC/PPCInstrPrefix.td

2201

This COPY_TO_REGCLASS and below are redundant and could cause issues in the future. Fixed in e73f885c988d.

llvm/lib/Target/PowerPC/PPCInstrPrefix.td

 
 let Predicates = [IsISA3_1, HasVSX] in {
   def : Pat<(v16i8 (int_ppc_vsx_xvcvspbf16 v16i8:$XA)),
             (COPY_TO_REGCLASS (XVCVSPBF16 RCCp.AToVSRC), VRRC)>;
   def : Pat<(v16i8 (int_ppc_vsx_xvcvbf16spn v16i8:$XA)),
             (COPY_TO_REGCLASS (XVCVBF16SPN RCCp.AToVSRC), VRRC)>;
 }
 
-let AddedComplexity = 400, Predicates = [IsISA3_1] in {
-  def : Pat<(truncstorei8 (i32 (vector_extract v16i8:$rS, 0)), xoaddr:$src),
-            (STXVRBX (COPY_TO_REGCLASS $rS, VSRC), xoaddr:$src)>;
-  def : Pat<(truncstorei16 (i32 (vector_extract v8i16:$rS, 0)), xoaddr:$src),
-            (STXVRHX (COPY_TO_REGCLASS $rS, VSRC), xoaddr:$src)>;
-  def : Pat<(store (i32 (vector_extract v4i32:$rS, 0)), xoaddr:$src),
-            (STXVRWX (COPY_TO_REGCLASS $rS, VSRC), xoaddr:$src)>;
-  def : Pat<(store (i64 (vector_extract v2i64:$rS, 0)), xoaddr:$src),
-            (STXVRDX (COPY_TO_REGCLASS $rS, VSRC), xoaddr:$src)>;
+let AddedComplexity = 400, Predicates = [IsISA3_1, IsLittleEndian] in {
+  // Store element 0 of a VSX register to memory
+  def : Pat<(truncstorei8 (i32 (vector_extract v16i8:$src, 0)), xoaddr:$dst),
+            (STXVRBX (COPY_TO_REGCLASS v16i8:$src, VSRC), xoaddr:$dst)>;
+  def : Pat<(truncstorei16 (i32 (vector_extract v8i16:$src, 0)), xoaddr:$dst),
+            (STXVRHX (COPY_TO_REGCLASS v8i16:$src, VSRC), xoaddr:$dst)>;
+  def : Pat<(store (i32 (extractelt v4i32:$src, 0)), xoaddr:$dst),
+            (STXVRWX (COPY_TO_REGCLASS v4i32:$src, VSRC), xoaddr:$dst)>;
+  def : Pat<(store (f32 (extractelt v4f32:$src, 0)), xoaddr:$dst),
+            (STXVRWX (COPY_TO_REGCLASS v4f32:$src, VSRC), xoaddr:$dst)>;
+  def : Pat<(store (i64 (extractelt v2i64:$src, 0)), xoaddr:$dst),
+            (STXVRDX (COPY_TO_REGCLASS v2i64:$src, VSRC), xoaddr:$dst)>;
+  def : Pat<(store (f64 (extractelt v2f64:$src, 0)), xoaddr:$dst),
+            (STXVRDX (COPY_TO_REGCLASS v2f64:$src, VSRC), xoaddr:$dst)>;
 }
 
 let AddedComplexity = 400, Predicates = [PrefixInstrs] in {
  def : Pat<(v4i32 (build_vector i32immNonAllOneNonZero:$A,
                                 i32immNonAllOneNonZero:$A,
                                 i32immNonAllOneNonZero:$A,
                                 i32immNonAllOneNonZero:$A)),
            (v4i32 (XXSPLTIW imm:$A))>;
  def : Pat<(f32 nzFPImmAsi32:$A),
            (COPY_TO_REGCLASS (XXSPLTIDP (getFPAs32BitInt fpimm:$A)),
                              VSFRC)>;
  def : Pat<(f64 nzFPImmAsi32:$A),
            (COPY_TO_REGCLASS (XXSPLTIDP (getFPAs32BitInt fpimm:$A)),
                              VSFRC)>;
 }
 

llvm/test/CodeGen/PowerPC/builtins-ppc-p10vsx.ll

 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu \
 ; RUN:   -mcpu=pwr10 -ppc-asm-full-reg-names -ppc-vsr-nums-as-vr < %s | \
-; RUN:   FileCheck %s
+; RUN:   FileCheck %s --check-prefix=CHECK-LE
 ; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-linux-gnu \
 ; RUN:   -mcpu=pwr10 -ppc-asm-full-reg-names -ppc-vsr-nums-as-vr < %s | \
-; RUN:   FileCheck %s
+; RUN:   FileCheck %s --check-prefix=CHECK-BE
 ; RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu -O0 \
 ; RUN:   -mcpu=pwr10 -ppc-asm-full-reg-names -ppc-vsr-nums-as-vr < %s | \
 ; RUN:   FileCheck %s --check-prefix=CHECK-O0
 
 ; These test cases aims to test the builtins for the Power10 VSX vector
 ; instructions introduced in ISA 3.1.
 
 declare i32 @llvm.ppc.vsx.xvtlsbb(<16 x i8>, i32)
 
 define signext i32 @test_vec_test_lsbb_all_ones(<16 x i8> %vuca) {
 ; CHECK-LABEL: test_vec_test_lsbb_all_ones:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    xvtlsbb cr0, v2
 ; CHECK-NEXT:    mfocrf r3, 128
 ; CHECK-NEXT:    srwi r3, r3, 31
 ; CHECK-NEXT:    extsw r3, r3
 ; CHECK-NEXT:    blr
 ;
+; CHECK-LE-LABEL: test_vec_test_lsbb_all_ones:
+; CHECK-LE:       # %bb.0: # %entry
+; CHECK-LE-NEXT:    xvtlsbb cr0, v2
+; CHECK-LE-NEXT:    mfocrf r3, 128
+; CHECK-LE-NEXT:    srwi r3, r3, 31
+; CHECK-LE-NEXT:    extsw r3, r3
+; CHECK-LE-NEXT:    blr
+;
+; CHECK-BE-LABEL: test_vec_test_lsbb_all_ones:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    xvtlsbb cr0, v2
+; CHECK-BE-NEXT:    mfocrf r3, 128
+; CHECK-BE-NEXT:    srwi r3, r3, 31
+; CHECK-BE-NEXT:    extsw r3, r3
+; CHECK-BE-NEXT:    blr
+;
 ; CHECK-O0-LABEL: test_vec_test_lsbb_all_ones:
 ; CHECK-O0:       # %bb.0: # %entry
 ; CHECK-O0-NEXT:    xvtlsbb cr0, v2
 ; CHECK-O0-NEXT:    mfocrf r3, 128
 ; CHECK-O0-NEXT:    srwi r3, r3, 31
 ; CHECK-O0-NEXT:    extsw r3, r3
 ; CHECK-O0-NEXT:    blr
 entry:
   %0 = tail call i32 @llvm.ppc.vsx.xvtlsbb(<16 x i8> %vuca, i32 1)
   ret i32 %0
 }
 
 define signext i32 @test_vec_test_lsbb_all_zeros(<16 x i8> %vuca) {
 ; CHECK-LABEL: test_vec_test_lsbb_all_zeros:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    xvtlsbb cr0, v2
 ; CHECK-NEXT:    mfocrf r3, 128
 ; CHECK-NEXT:    rlwinm r3, r3, 3, 31, 31
 ; CHECK-NEXT:    extsw r3, r3
 ; CHECK-NEXT:    blr
 ;
+; CHECK-LE-LABEL: test_vec_test_lsbb_all_zeros:
+; CHECK-LE:       # %bb.0: # %entry
+; CHECK-LE-NEXT:    xvtlsbb cr0, v2
+; CHECK-LE-NEXT:    mfocrf r3, 128
+; CHECK-LE-NEXT:    rlwinm r3, r3, 3, 31, 31
+; CHECK-LE-NEXT:    extsw r3, r3
+; CHECK-LE-NEXT:    blr
+;
+; CHECK-BE-LABEL: test_vec_test_lsbb_all_zeros:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    xvtlsbb cr0, v2
+; CHECK-BE-NEXT:    mfocrf r3, 128
+; CHECK-BE-NEXT:    rlwinm r3, r3, 3, 31, 31
+; CHECK-BE-NEXT:    extsw r3, r3
+; CHECK-BE-NEXT:    blr
+;
 ; CHECK-O0-LABEL: test_vec_test_lsbb_all_zeros:
 ; CHECK-O0:       # %bb.0: # %entry
 ; CHECK-O0-NEXT:    xvtlsbb cr0, v2
 ; CHECK-O0-NEXT:    mfocrf r3, 128
 ; CHECK-O0-NEXT:    rlwinm r3, r3, 3, 31, 31
 ; CHECK-O0-NEXT:    extsw r3, r3
 ; CHECK-O0-NEXT:    blr
 entry:
   %0 = tail call i32 @llvm.ppc.vsx.xvtlsbb(<16 x i8> %vuca, i32 0)
   ret i32 %0
 }
 
 define void @vec_xst_trunc_sc(<1 x i128> %__vec, i64 %__offset, i8* nocapture %__ptr) {
 ; CHECK-LABEL: vec_xst_trunc_sc:
 ; CHECK:       # %bb.0: # %entry
-; CHECK-NEXT:    stxvrbx v2, r6, r5
+; CHECK-NEXT:    vsldoi v2, v2, v2, 9
+; CHECK-NEXT:    stxsibx v2, r6, r5
 ; CHECK-NEXT:    blr
 ;
+; CHECK-LE-LABEL: vec_xst_trunc_sc:
+; CHECK-LE:       # %bb.0: # %entry
+; CHECK-LE-NEXT:    stxvrbx v2, r6, r5
+; CHECK-LE-NEXT:    blr
+;
+; CHECK-BE-LABEL: vec_xst_trunc_sc:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    vsldoi v2, v2, v2, 9
+; CHECK-BE-NEXT:    stxsibx v2, r6, r5
+; CHECK-BE-NEXT:    blr
+;
 ; CHECK-O0-LABEL: vec_xst_trunc_sc:
 ; CHECK-O0:       # %bb.0: # %entry
 ; CHECK-O0-NEXT:    li r3, 0
 ; CHECK-O0-NEXT:    vextubrx r3, r3, v2
 ; CHECK-O0-NEXT:    # kill: def $r3 killed $r3 killed $x3
 ; CHECK-O0-NEXT:    add r4, r6, r5
 ; CHECK-O0-NEXT:    stb r3, 0(r4)
 ; CHECK-O0-NEXT:    blr
 entry:
   %0 = bitcast <1 x i128> %__vec to <16 x i8>
   %conv = extractelement <16 x i8> %0, i32 0
   %add.ptr = getelementptr inbounds i8, i8* %__ptr, i64 %__offset
   store i8 %conv, i8* %add.ptr, align 1
   ret void
 }
 
 define void @vec_xst_trunc_uc(<1 x i128> %__vec, i64 %__offset, i8* nocapture %__ptr) {
 ; CHECK-LABEL: vec_xst_trunc_uc:
 ; CHECK:       # %bb.0: # %entry
-; CHECK-NEXT:    stxvrbx v2, r6, r5
+; CHECK-NEXT:    vsldoi v2, v2, v2, 9
+; CHECK-NEXT:    stxsibx v2, r6, r5
 ; CHECK-NEXT:    blr
 ;
+; CHECK-LE-LABEL: vec_xst_trunc_uc:
+; CHECK-LE:       # %bb.0: # %entry
+; CHECK-LE-NEXT:    stxvrbx v2, r6, r5
+; CHECK-LE-NEXT:    blr
+;
+; CHECK-BE-LABEL: vec_xst_trunc_uc:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    vsldoi v2, v2, v2, 9
+; CHECK-BE-NEXT:    stxsibx v2, r6, r5
+; CHECK-BE-NEXT:    blr
+;
 ; CHECK-O0-LABEL: vec_xst_trunc_uc:
 ; CHECK-O0:       # %bb.0: # %entry
 ; CHECK-O0-NEXT:    li r3, 0
 ; CHECK-O0-NEXT:    vextubrx r3, r3, v2
 ; CHECK-O0-NEXT:    # kill: def $r3 killed $r3 killed $x3
 ; CHECK-O0-NEXT:    add r4, r6, r5
 ; CHECK-O0-NEXT:    stb r3, 0(r4)
 ; CHECK-O0-NEXT:    blr
 entry:
   %0 = bitcast <1 x i128> %__vec to <16 x i8>
   %conv = extractelement <16 x i8> %0, i32 0
   %add.ptr = getelementptr inbounds i8, i8* %__ptr, i64 %__offset
   store i8 %conv, i8* %add.ptr, align 1
   ret void
 }
 
 define void @vec_xst_trunc_ss(<1 x i128> %__vec, i64 %__offset, i16* nocapture %__ptr) {
 ; CHECK-LABEL: vec_xst_trunc_ss:
 ; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsldoi v2, v2, v2, 10
 ; CHECK-NEXT:    sldi r3, r5, 1
-; CHECK-NEXT:    stxvrhx v2, r6, r3
+; CHECK-NEXT:    stxsihx v2, r6, r3
 ; CHECK-NEXT:    blr
 ;
+; CHECK-LE-LABEL: vec_xst_trunc_ss:
+; CHECK-LE:       # %bb.0: # %entry
+; CHECK-LE-NEXT:    sldi r3, r5, 1
+; CHECK-LE-NEXT:    stxvrhx v2, r6, r3
+; CHECK-LE-NEXT:    blr
+;
+; CHECK-BE-LABEL: vec_xst_trunc_ss:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    vsldoi v2, v2, v2, 10
+; CHECK-BE-NEXT:    sldi r3, r5, 1
+; CHECK-BE-NEXT:    stxsihx v2, r6, r3
+; CHECK-BE-NEXT:    blr
+;
 ; CHECK-O0-LABEL: vec_xst_trunc_ss:
 ; CHECK-O0:       # %bb.0: # %entry
 ; CHECK-O0-NEXT:    li r3, 0
 ; CHECK-O0-NEXT:    vextuhrx r3, r3, v2
 ; CHECK-O0-NEXT:    # kill: def $r3 killed $r3 killed $x3
 ; CHECK-O0-NEXT:    sldi r4, r5, 1
 ; CHECK-O0-NEXT:    add r4, r6, r4
 ; CHECK-O0-NEXT:    sth r3, 0(r4)
 ; CHECK-O0-NEXT:    blr
 entry:
   %0 = bitcast <1 x i128> %__vec to <8 x i16>
   %conv = extractelement <8 x i16> %0, i32 0
   %add.ptr = getelementptr inbounds i16, i16* %__ptr, i64 %__offset
   store i16 %conv, i16* %add.ptr, align 2
   ret void
 }
 
 define void @vec_xst_trunc_us(<1 x i128> %__vec, i64 %__offset, i16* nocapture %__ptr) {
 ; CHECK-LABEL: vec_xst_trunc_us:
 ; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsldoi v2, v2, v2, 10
 ; CHECK-NEXT:    sldi r3, r5, 1
-; CHECK-NEXT:    stxvrhx v2, r6, r3
+; CHECK-NEXT:    stxsihx v2, r6, r3
 ; CHECK-NEXT:    blr
 ;
+; CHECK-LE-LABEL: vec_xst_trunc_us:
+; CHECK-LE:       # %bb.0: # %entry
+; CHECK-LE-NEXT:    sldi r3, r5, 1
+; CHECK-LE-NEXT:    stxvrhx v2, r6, r3
+; CHECK-LE-NEXT:    blr
+;
+; CHECK-BE-LABEL: vec_xst_trunc_us:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    vsldoi v2, v2, v2, 10
+; CHECK-BE-NEXT:    sldi r3, r5, 1
+; CHECK-BE-NEXT:    stxsihx v2, r6, r3
+; CHECK-BE-NEXT:    blr
+;
 ; CHECK-O0-LABEL: vec_xst_trunc_us:
 ; CHECK-O0:       # %bb.0: # %entry
 ; CHECK-O0-NEXT:    li r3, 0
 ; CHECK-O0-NEXT:    vextuhrx r3, r3, v2
 ; CHECK-O0-NEXT:    # kill: def $r3 killed $r3 killed $x3
 ; CHECK-O0-NEXT:    sldi r4, r5, 1
 ; CHECK-O0-NEXT:    add r4, r6, r4
 ; CHECK-O0-NEXT:    sth r3, 0(r4)
 ; CHECK-O0-NEXT:    blr
 entry:
   %0 = bitcast <1 x i128> %__vec to <8 x i16>
   %conv = extractelement <8 x i16> %0, i32 0
   %add.ptr = getelementptr inbounds i16, i16* %__ptr, i64 %__offset
   store i16 %conv, i16* %add.ptr, align 2
   ret void
 }
 
 define void @vec_xst_trunc_si(<1 x i128> %__vec, i64 %__offset, i32* nocapture %__ptr) {
 ; CHECK-LABEL: vec_xst_trunc_si:
 ; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    xxsldwi vs0, v2, v2, 3
 ; CHECK-NEXT:    sldi r3, r5, 2
-; CHECK-NEXT:    stxvrwx v2, r6, r3
+; CHECK-NEXT:    stfiwx f0, r6, r3
 ; CHECK-NEXT:    blr
 ;
+; CHECK-LE-LABEL: vec_xst_trunc_si:
+; CHECK-LE:       # %bb.0: # %entry
+; CHECK-LE-NEXT:    sldi r3, r5, 2
+; CHECK-LE-NEXT:    stxvrwx v2, r6, r3
+; CHECK-LE-NEXT:    blr
+;
+; CHECK-BE-LABEL: vec_xst_trunc_si:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    xxsldwi vs0, v2, v2, 3
+; CHECK-BE-NEXT:    sldi r3, r5, 2
+; CHECK-BE-NEXT:    stfiwx f0, r6, r3
+; CHECK-BE-NEXT:    blr
+;
 ; CHECK-O0-LABEL: vec_xst_trunc_si:
 ; CHECK-O0:       # %bb.0: # %entry
 ; CHECK-O0-NEXT:    li r3, 0
 ; CHECK-O0-NEXT:    vextuwrx r3, r3, v2
 ; CHECK-O0-NEXT:    # kill: def $r3 killed $r3 killed $x3
 ; CHECK-O0-NEXT:    sldi r4, r5, 2
 ; CHECK-O0-NEXT:    add r4, r6, r4
 ; CHECK-O0-NEXT:    stw r3, 0(r4)
 ; CHECK-O0-NEXT:    blr
 entry:
   %0 = bitcast <1 x i128> %__vec to <4 x i32>
   %conv = extractelement <4 x i32> %0, i32 0
   %add.ptr = getelementptr inbounds i32, i32* %__ptr, i64 %__offset
   store i32 %conv, i32* %add.ptr, align 4
   ret void
 }
 
 define void @vec_xst_trunc_ui(<1 x i128> %__vec, i64 %__offset, i32* nocapture %__ptr) {
 ; CHECK-LABEL: vec_xst_trunc_ui:
 ; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    xxsldwi vs0, v2, v2, 3
 ; CHECK-NEXT:    sldi r3, r5, 2
-; CHECK-NEXT:    stxvrwx v2, r6, r3
+; CHECK-NEXT:    stfiwx f0, r6, r3
 ; CHECK-NEXT:    blr
 ;
+; CHECK-LE-LABEL: vec_xst_trunc_ui:
+; CHECK-LE:       # %bb.0: # %entry
+; CHECK-LE-NEXT:    sldi r3, r5, 2
+; CHECK-LE-NEXT:    stxvrwx v2, r6, r3
+; CHECK-LE-NEXT:    blr
+;
+; CHECK-BE-LABEL: vec_xst_trunc_ui:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    xxsldwi vs0, v2, v2, 3
+; CHECK-BE-NEXT:    sldi r3, r5, 2
+; CHECK-BE-NEXT:    stfiwx f0, r6, r3
+; CHECK-BE-NEXT:    blr
+;
 ; CHECK-O0-LABEL: vec_xst_trunc_ui:
 ; CHECK-O0:       # %bb.0: # %entry
 ; CHECK-O0-NEXT:    li r3, 0
 ; CHECK-O0-NEXT:    vextuwrx r3, r3, v2
 ; CHECK-O0-NEXT:    # kill: def $r3 killed $r3 killed $x3
 ; CHECK-O0-NEXT:    sldi r4, r5, 2
 ; CHECK-O0-NEXT:    add r4, r6, r4
 ; CHECK-O0-NEXT:    stw r3, 0(r4)
 ; CHECK-O0-NEXT:    blr
 entry:
   %0 = bitcast <1 x i128> %__vec to <4 x i32>
   %conv = extractelement <4 x i32> %0, i32 0
   %add.ptr = getelementptr inbounds i32, i32* %__ptr, i64 %__offset
   store i32 %conv, i32* %add.ptr, align 4
   ret void
 }
 
 define void @vec_xst_trunc_sll(<1 x i128> %__vec, i64 %__offset, i64* nocapture %__ptr)  {
 ; CHECK-LABEL: vec_xst_trunc_sll:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    sldi r3, r5, 3
-; CHECK-NEXT:    stxvrdx v2, r6, r3
+; CHECK-NEXT:    stxsdx v2, r6, r3
 ; CHECK-NEXT:    blr
 ;
+; CHECK-LE-LABEL: vec_xst_trunc_sll:
+; CHECK-LE:       # %bb.0: # %entry
+; CHECK-LE-NEXT:    sldi r3, r5, 3
+; CHECK-LE-NEXT:    stxvrdx v2, r6, r3
+; CHECK-LE-NEXT:    blr
+;
+; CHECK-BE-LABEL: vec_xst_trunc_sll:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    sldi r3, r5, 3
+; CHECK-BE-NEXT:    stxsdx v2, r6, r3
+; CHECK-BE-NEXT:    blr
+;
 ; CHECK-O0-LABEL: vec_xst_trunc_sll:
 ; CHECK-O0:       # %bb.0: # %entry
 ; CHECK-O0-NEXT:    mfvsrld r3, v2
 ; CHECK-O0-NEXT:    sldi r4, r5, 3
 ; CHECK-O0-NEXT:    add r4, r6, r4
 ; CHECK-O0-NEXT:    std r3, 0(r4)
 ; CHECK-O0-NEXT:    blr
 entry:
   %0 = bitcast <1 x i128> %__vec to <2 x i64>
   %conv = extractelement <2 x i64> %0, i32 0
   %add.ptr = getelementptr inbounds i64, i64* %__ptr, i64 %__offset
   store i64 %conv, i64* %add.ptr, align 8
   ret void
 }
 
 define void @vec_xst_trunc_ull(<1 x i128> %__vec, i64 %__offset, i64* nocapture %__ptr)  {
 ; CHECK-LABEL: vec_xst_trunc_ull:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    sldi r3, r5, 3
-; CHECK-NEXT:    stxvrdx v2, r6, r3
+; CHECK-NEXT:    stxsdx v2, r6, r3
 ; CHECK-NEXT:    blr
 ;
+; CHECK-LE-LABEL: vec_xst_trunc_ull:
+; CHECK-LE:       # %bb.0: # %entry
+; CHECK-LE-NEXT:    sldi r3, r5, 3
+; CHECK-LE-NEXT:    stxvrdx v2, r6, r3
+; CHECK-LE-NEXT:    blr
+;
+; CHECK-BE-LABEL: vec_xst_trunc_ull:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    sldi r3, r5, 3
+; CHECK-BE-NEXT:    stxsdx v2, r6, r3
+; CHECK-BE-NEXT:    blr
+;
 ; CHECK-O0-LABEL: vec_xst_trunc_ull:
 ; CHECK-O0:       # %bb.0: # %entry
 ; CHECK-O0-NEXT:    mfvsrld r3, v2
 ; CHECK-O0-NEXT:    sldi r4, r5, 3
 ; CHECK-O0-NEXT:    add r4, r6, r4
 ; CHECK-O0-NEXT:    std r3, 0(r4)
 ; CHECK-O0-NEXT:    blr
 entry:
   %0 = bitcast <1 x i128> %__vec to <2 x i64>
   %conv = extractelement <2 x i64> %0, i32 0
   %add.ptr = getelementptr inbounds i64, i64* %__ptr, i64 %__offset
   store i64 %conv, i64* %add.ptr, align 8
   ret void
 }
 
 define dso_local <1 x i128> @vec_xl_zext(i64 %__offset, i8* nocapture readonly %__pointer) {
 ; CHECK-LABEL: vec_xl_zext:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    lxvrbx v2, r4, r3
 ; CHECK-NEXT:    blr
 ;
+; CHECK-LE-LABEL: vec_xl_zext:
+; CHECK-LE:       # %bb.0: # %entry
+; CHECK-LE-NEXT:    lxvrbx v2, r4, r3
+; CHECK-LE-NEXT:    blr
+;
+; CHECK-BE-LABEL: vec_xl_zext:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lxvrbx v2, r4, r3
+; CHECK-BE-NEXT:    blr
+;
 ; CHECK-O0-LABEL: vec_xl_zext:
 ; CHECK-O0:       # %bb.0: # %entry
 ; CHECK-O0-NEXT:    lxvrbx v2, r4, r3
 ; CHECK-O0-NEXT:    blr
 entry:
   %add.ptr = getelementptr inbounds i8, i8* %__pointer, i64 %__offset
   %0 = load i8, i8* %add.ptr, align 1
   %conv = zext i8 %0 to i128
   %splat.splatinsert = insertelement <1 x i128> undef, i128 %conv, i32 0
   ret <1 x i128> %splat.splatinsert
 }
 
 define dso_local <1 x i128> @vec_xl_zext_short(i64 %__offset, i16* nocapture readonly %__pointer) {
 ; CHECK-LABEL: vec_xl_zext_short:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    sldi r3, r3, 1
 ; CHECK-NEXT:    lxvrhx v2, r4, r3
 ; CHECK-NEXT:    blr
 ;
+; CHECK-LE-LABEL: vec_xl_zext_short:
+; CHECK-LE:       # %bb.0: # %entry
+; CHECK-LE-NEXT:    sldi r3, r3, 1
+; CHECK-LE-NEXT:    lxvrhx v2, r4, r3
+; CHECK-LE-NEXT:    blr
+;
+; CHECK-BE-LABEL: vec_xl_zext_short:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    sldi r3, r3, 1
+; CHECK-BE-NEXT:    lxvrhx v2, r4, r3
+; CHECK-BE-NEXT:    blr
+;
 ; CHECK-O0-LABEL: vec_xl_zext_short:
 ; CHECK-O0:       # %bb.0: # %entry
 ; CHECK-O0-NEXT:    sldi r3, r3, 1
 ; CHECK-O0-NEXT:    lxvrhx v2, r4, r3
 ; CHECK-O0-NEXT:    blr
 entry:
   %add.ptr = getelementptr inbounds i16, i16* %__pointer, i64 %__offset
   %0 = load i16, i16* %add.ptr, align 2
   %conv = zext i16 %0 to i128
   %splat.splatinsert = insertelement <1 x i128> undef, i128 %conv, i32 0
   ret <1 x i128> %splat.splatinsert
 }
 
 define dso_local <1 x i128> @vec_xl_zext_word(i64 %__offset, i32* nocapture readonly %__pointer) {
 ; CHECK-LABEL: vec_xl_zext_word:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    sldi r3, r3, 2
 ; CHECK-NEXT:    lxvrwx v2, r4, r3
 ; CHECK-NEXT:    blr
 ;
+; CHECK-LE-LABEL: vec_xl_zext_word:
+; CHECK-LE:       # %bb.0: # %entry
+; CHECK-LE-NEXT:    sldi r3, r3, 2
+; CHECK-LE-NEXT:    lxvrwx v2, r4, r3
+; CHECK-LE-NEXT:    blr
+;
+; CHECK-BE-LABEL: vec_xl_zext_word:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    sldi r3, r3, 2
+; CHECK-BE-NEXT:    lxvrwx v2, r4, r3
+; CHECK-BE-NEXT:    blr
+;
 ; CHECK-O0-LABEL: vec_xl_zext_word:
 ; CHECK-O0:       # %bb.0: # %entry
 ; CHECK-O0-NEXT:    sldi r3, r3, 2
 ; CHECK-O0-NEXT:    lxvrwx v2, r4, r3
 ; CHECK-O0-NEXT:    blr
 entry:
   %add.ptr = getelementptr inbounds i32, i32* %__pointer, i64 %__offset
   %0 = load i32, i32* %add.ptr, align 4
   %conv = zext i32 %0 to i128
   %splat.splatinsert = insertelement <1 x i128> undef, i128 %conv, i32 0
   ret <1 x i128> %splat.splatinsert
 }
 
 define dso_local <1 x i128> @vec_xl_zext_dw(i64 %__offset, i64* nocapture readonly %__pointer) {
 ; CHECK-LABEL: vec_xl_zext_dw:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    sldi r3, r3, 3
 ; CHECK-NEXT:    lxvrdx v2, r4, r3
 ; CHECK-NEXT:    blr
 ;
+; CHECK-LE-LABEL: vec_xl_zext_dw:
+; CHECK-LE:       # %bb.0: # %entry
+; CHECK-LE-NEXT:    sldi r3, r3, 3
+; CHECK-LE-NEXT:    lxvrdx v2, r4, r3
+; CHECK-LE-NEXT:    blr
+;
+; CHECK-BE-LABEL: vec_xl_zext_dw:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    sldi r3, r3, 3
+; CHECK-BE-NEXT:    lxvrdx v2, r4, r3
+; CHECK-BE-NEXT:    blr
+;
 ; CHECK-O0-LABEL: vec_xl_zext_dw:
 ; CHECK-O0:       # %bb.0: # %entry
 ; CHECK-O0-NEXT:    sldi r3, r3, 3
 ; CHECK-O0-NEXT:    lxvrdx v2, r4, r3
 ; CHECK-O0-NEXT:    blr
 entry:
   %add.ptr = getelementptr inbounds i64, i64* %__pointer, i64 %__offset
   %0 = load i64, i64* %add.ptr, align 8
   %conv = zext i64 %0 to i128
   %splat.splatinsert = insertelement <1 x i128> undef, i128 %conv, i32 0
   ret <1 x i128> %splat.splatinsert
 }
 
 define dso_local <1 x i128> @vec_xl_sext_b(i64 %offset, i8* %p) {
 ; CHECK-LABEL: vec_xl_sext_b:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    lbzx r3, r4, r3
 ; CHECK-NEXT:    extsb r3, r3
 ; CHECK-NEXT:    sradi r4, r3, 63
 ; CHECK-NEXT:    mtvsrdd v2, r4, r3
 ; CHECK-NEXT:    blr
 ;
+; CHECK-LE-LABEL: vec_xl_sext_b:
+; CHECK-LE:       # %bb.0: # %entry
+; CHECK-LE-NEXT:    lbzx r3, r4, r3
+; CHECK-LE-NEXT:    extsb r3, r3
+; CHECK-LE-NEXT:    sradi r4, r3, 63
+; CHECK-LE-NEXT:    mtvsrdd v2, r4, r3
+; CHECK-LE-NEXT:    blr
+;
+; CHECK-BE-LABEL: vec_xl_sext_b:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    lbzx r3, r4, r3
+; CHECK-BE-NEXT:    extsb r3, r3
+; CHECK-BE-NEXT:    sradi r4, r3, 63
+; CHECK-BE-NEXT:    mtvsrdd v2, r4, r3
+; CHECK-BE-NEXT:    blr
+;
 ; CHECK-O0-LABEL: vec_xl_sext_b:
 ; CHECK-O0:       # %bb.0: # %entry
 ; CHECK-O0-NEXT:    lbzx r3, r4, r3
 ; CHECK-O0-NEXT:    extsb r4, r3
 ; CHECK-O0-NEXT:    sradi r3, r4, 63
 ; CHECK-O0-NEXT:    mtvsrdd v2, r3, r4
 ; CHECK-O0-NEXT:    blr
 entry:
   %add.ptr = getelementptr inbounds i8, i8* %p, i64 %offset
   %0 = load i8, i8* %add.ptr, align 1
   %conv = sext i8 %0 to i128
   %splat.splatinsert = insertelement <1 x i128> undef, i128 %conv, i32 0
   ret <1 x i128> %splat.splatinsert
 }
 
 define dso_local <1 x i128> @vec_xl_sext_h(i64 %offset, i16* %p) {
 ; CHECK-LABEL: vec_xl_sext_h:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    sldi r3, r3, 1
 ; CHECK-NEXT:    lhax r3, r4, r3
 ; CHECK-NEXT:    sradi r4, r3, 63
 ; CHECK-NEXT:    mtvsrdd v2, r4, r3
 ; CHECK-NEXT:    blr
 ;
+; CHECK-LE-LABEL: vec_xl_sext_h:
+; CHECK-LE:       # %bb.0: # %entry
+; CHECK-LE-NEXT:    sldi r3, r3, 1
+; CHECK-LE-NEXT:    lhax r3, r4, r3
+; CHECK-LE-NEXT:    sradi r4, r3, 63
+; CHECK-LE-NEXT:    mtvsrdd v2, r4, r3
+; CHECK-LE-NEXT:    blr
+;
+; CHECK-BE-LABEL: vec_xl_sext_h:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    sldi r3, r3, 1
+; CHECK-BE-NEXT:    lhax r3, r4, r3
+; CHECK-BE-NEXT:    sradi r4, r3, 63
+; CHECK-BE-NEXT:    mtvsrdd v2, r4, r3
+; CHECK-BE-NEXT:    blr
+;
 ; CHECK-O0-LABEL: vec_xl_sext_h:
 ; CHECK-O0:       # %bb.0: # %entry
 ; CHECK-O0-NEXT:    sldi r3, r3, 1
 ; CHECK-O0-NEXT:    lhax r4, r4, r3
 ; CHECK-O0-NEXT:    sradi r3, r4, 63
 ; CHECK-O0-NEXT:    mtvsrdd v2, r3, r4
 ; CHECK-O0-NEXT:    blr
 entry:
   %add.ptr = getelementptr inbounds i16, i16* %p, i64 %offset
   %0 = load i16, i16* %add.ptr, align 2
   %conv = sext i16 %0 to i128
   %splat.splatinsert = insertelement <1 x i128> undef, i128 %conv, i32 0
   ret <1 x i128> %splat.splatinsert
 }
 
 define dso_local <1 x i128> @vec_xl_sext_w(i64 %offset, i32* %p) {
 ; CHECK-LABEL: vec_xl_sext_w:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    sldi r3, r3, 2
 ; CHECK-NEXT:    lwax r3, r4, r3
 ; CHECK-NEXT:    sradi r4, r3, 63
 ; CHECK-NEXT:    mtvsrdd v2, r4, r3
 ; CHECK-NEXT:    blr
 ;
+; CHECK-LE-LABEL: vec_xl_sext_w:
+; CHECK-LE:       # %bb.0: # %entry
+; CHECK-LE-NEXT:    sldi r3, r3, 2
+; CHECK-LE-NEXT:    lwax r3, r4, r3
+; CHECK-LE-NEXT:    sradi r4, r3, 63
+; CHECK-LE-NEXT:    mtvsrdd v2, r4, r3
+; CHECK-LE-NEXT:    blr
+;
+; CHECK-BE-LABEL: vec_xl_sext_w:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    sldi r3, r3, 2
+; CHECK-BE-NEXT:    lwax r3, r4, r3
+; CHECK-BE-NEXT:    sradi r4, r3, 63
+; CHECK-BE-NEXT:    mtvsrdd v2, r4, r3
+; CHECK-BE-NEXT:    blr
+;
 ; CHECK-O0-LABEL: vec_xl_sext_w:
 ; CHECK-O0:       # %bb.0: # %entry
 ; CHECK-O0-NEXT:    sldi r3, r3, 2
 ; CHECK-O0-NEXT:    lwax r4, r4, r3
 ; CHECK-O0-NEXT:    sradi r3, r4, 63
 ; CHECK-O0-NEXT:    mtvsrdd v2, r3, r4
 ; CHECK-O0-NEXT:    blr
 entry:
   %add.ptr = getelementptr inbounds i32, i32* %p, i64 %offset
   %0 = load i32, i32* %add.ptr, align 4
   %conv = sext i32 %0 to i128
   %splat.splatinsert = insertelement <1 x i128> undef, i128 %conv, i32 0
   ret <1 x i128> %splat.splatinsert
 }
 
 define dso_local <1 x i128> @vec_xl_sext_d(i64 %offset, i64* %p) {
 ; CHECK-LABEL: vec_xl_sext_d:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    sldi r3, r3, 3
 ; CHECK-NEXT:    ldx r3, r4, r3
 ; CHECK-NEXT:    sradi r4, r3, 63
 ; CHECK-NEXT:    mtvsrdd v2, r4, r3
 ; CHECK-NEXT:    blr
 ;
+; CHECK-LE-LABEL: vec_xl_sext_d:
+; CHECK-LE:       # %bb.0: # %entry
+; CHECK-LE-NEXT:    sldi r3, r3, 3
+; CHECK-LE-NEXT:    ldx r3, r4, r3
+; CHECK-LE-NEXT:    sradi r4, r3, 63
+; CHECK-LE-NEXT:    mtvsrdd v2, r4, r3
+; CHECK-LE-NEXT:    blr
+;
+; CHECK-BE-LABEL: vec_xl_sext_d:
+; CHECK-BE:       # %bb.0: # %entry
+; CHECK-BE-NEXT:    sldi r3, r3, 3
+; CHECK-BE-NEXT:    ldx r3, r4, r3
+; CHECK-BE-NEXT:    sradi r4, r3, 63
+; CHECK-BE-NEXT:    mtvsrdd v2, r4, r3
+; CHECK-BE-NEXT:    blr
+;
 ; CHECK-O0-LABEL: vec_xl_sext_d:
 ; CHECK-O0:       # %bb.0: # %entry
 ; CHECK-O0-NEXT:    sldi r3, r3, 3
 ; CHECK-O0-NEXT:    ldx r4, r4, r3
 ; CHECK-O0-NEXT:    sradi r3, r4, 63
 ; CHECK-O0-NEXT:    mtvsrdd v2, r3, r4
 ; CHECK-O0-NEXT:    blr
 entry:
   %add.ptr = getelementptr inbounds i64, i64* %p, i64 %offset
   %0 = load i64, i64* %add.ptr, align 8
   %conv = sext i64 %0 to i128
   %splat.splatinsert = insertelement <1 x i128> undef, i128 %conv, i32 0
   ret <1 x i128> %splat.splatinsert
 }

llvm/test/CodeGen/PowerPC/store-rightmost-vector-elt.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu \
+; RUN:     -mcpu=pwr10 -ppc-vsr-nums-as-vr -ppc-asm-full-reg-names \
+; RUN:     < %s | FileCheck %s
+
+define void @test1(<4 x i32> %A, i32* %a) {
+; CHECK-LABEL: test1:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    stxvrwx v2, 0, r5
+; CHECK-NEXT:    blr
+entry:
+  %vecext = extractelement <4 x i32> %A, i32 0
+  store i32 %vecext, i32* %a, align 4
+  ret void
+}
+
+define void @test2(<4 x float> %A, float* %a) {
+; CHECK-LABEL: test2:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    stxvrwx v2, 0, r5
+; CHECK-NEXT:    blr
+entry:
+  %vecext = extractelement <4 x float> %A, i32 0
+  store float %vecext, float* %a, align 4
+  ret void
+}
+
+define void @test3(<2 x double> %A, double* %a) {
+; CHECK-LABEL: test3:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    stxvrdx v2, 0, r5
+; CHECK-NEXT:    blr
+entry:
+  %vecext = extractelement <2 x double> %A, i32 0
+  store double %vecext, double* %a, align 8
+  ret void
+}
+
+define void @test4(<2 x i64> %A, i64* %a) {
+; CHECK-LABEL: test4:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    stxvrdx v2, 0, r5
+; CHECK-NEXT:    blr
+entry:
+  %vecext = extractelement <2 x i64> %A, i32 0
+  store i64 %vecext, i64* %a, align 8
+  ret void
+}
+
+define void @test5(<8 x i16> %A, i16* %a) {
+; CHECK-LABEL: test5:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    stxvrhx v2, 0, r5
+; CHECK-NEXT:    blr
+entry:
+  %vecext = extractelement <8 x i16> %A, i32 0
+  store i16 %vecext, i16* %a, align 2
+  ret void
+}
+
+define void @test6(<16 x i8> %A, i8* %a) {
+; CHECK-LABEL: test6:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    stxvrbx v2, 0, r5
+; CHECK-NEXT:    blr
+entry:
+  %vecext = extractelement <16 x i8> %A, i32 0
+  store i8 %vecext, i8* %a, align 1
+  ret void
+}
+