[SVE] Add intrinsics for uniform dsp operations that explicitly undefine the result for inactive lanes.

[dewen]

Summary

This patch adds new intrinsics for uniform dsp operations and changes the lowering for the following builtins to emit calls to the new aarch64.sve.###.u intrinsics.

svsqsub_x
svsqsub_n_x
svuqsub_x
svuqsub_n_x
svsqsubr_x
svsqsubr_n_x
svuqsubr_x
svuqsubr_n_x

[paulwalker-arm]

One minor suggestion but otherwise looks good.

llvm/test/CodeGen/AArch64/sve2-intrinsics-uniform-dsp-undef.ll

14

Please align the function parameters when landing the patch.

[dewen]

In D144704#4150855, @paulwalker-arm wrote:

One minor suggestion but otherwise looks good.

Okay, thank you for your review. I'll align the patch.

clang/include/clang/Basic/arm_sve.td

 let TargetGuard = "sve2" in {
 defm SVQADD_S  : SInstZPZZ<"svqadd",  "csli",     "aarch64_sve_sqadd",  "aarch64_sve_sqadd">;
 defm SVQADD_U  : SInstZPZZ<"svqadd",  "UcUsUiUl", "aarch64_sve_uqadd",  "aarch64_sve_uqadd">;
 defm SVHADD_S  : SInstZPZZ<"svhadd",  "csli",     "aarch64_sve_shadd",  "aarch64_sve_shadd">;
 defm SVHADD_U  : SInstZPZZ<"svhadd",  "UcUsUiUl", "aarch64_sve_uhadd",  "aarch64_sve_uhadd">;
 defm SVRHADD_S : SInstZPZZ<"svrhadd", "csli",     "aarch64_sve_srhadd", "aarch64_sve_srhadd">;
 defm SVRHADD_U : SInstZPZZ<"svrhadd", "UcUsUiUl", "aarch64_sve_urhadd", "aarch64_sve_urhadd">;
 
-defm SVQSUB_S  : SInstZPZZ<"svqsub",  "csli",     "aarch64_sve_sqsub",  "aarch64_sve_sqsub">;
-defm SVQSUB_U  : SInstZPZZ<"svqsub",  "UcUsUiUl", "aarch64_sve_uqsub",  "aarch64_sve_uqsub">;
-defm SVQSUBR_S : SInstZPZZ<"svqsubr", "csli",     "aarch64_sve_sqsubr", "aarch64_sve_sqsubr">;
-defm SVQSUBR_U : SInstZPZZ<"svqsubr", "UcUsUiUl", "aarch64_sve_uqsubr", "aarch64_sve_uqsubr">;
+defm SVQSUB_S  : SInstZPZZ<"svqsub",  "csli",     "aarch64_sve_sqsub",  "aarch64_sve_sqsub_u">;
+defm SVQSUB_U  : SInstZPZZ<"svqsub",  "UcUsUiUl", "aarch64_sve_uqsub",  "aarch64_sve_uqsub_u">;
+defm SVQSUBR_S : SInstZPZZ<"svqsubr", "csli",     "aarch64_sve_sqsubr", "aarch64_sve_sqsub_u", [ReverseMergeAnyBinOp]>;
+defm SVQSUBR_U : SInstZPZZ<"svqsubr", "UcUsUiUl", "aarch64_sve_uqsubr", "aarch64_sve_uqsub_u", [ReverseMergeAnyBinOp]>;
 defm SVHSUB_S  : SInstZPZZ<"svhsub",  "csli",     "aarch64_sve_shsub",  "aarch64_sve_shsub">;
 defm SVHSUB_U  : SInstZPZZ<"svhsub",  "UcUsUiUl", "aarch64_sve_uhsub",  "aarch64_sve_uhsub">;
 defm SVHSUBR_S : SInstZPZZ<"svhsubr", "csli",     "aarch64_sve_shsubr", "aarch64_sve_shsubr">;
 defm SVHSUBR_U : SInstZPZZ<"svhsubr", "UcUsUiUl", "aarch64_sve_uhsubr", "aarch64_sve_uhsubr">;
 
 defm SVQABS   : SInstZPZ<"svqabs",   "csil", "aarch64_sve_sqabs">;
 defm SVQNEG   : SInstZPZ<"svqneg",   "csil", "aarch64_sve_sqneg">;
 defm SVRECPE  : SInstZPZ<"svrecpe",  "Ui",   "aarch64_sve_urecpe">;

clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qsub.c

 //
 svuint64_t test_svqsub_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
   return SVE_ACLE_FUNC(svqsub,_u64,_m,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsub_s8_x(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.u.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z16test_svqsub_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.u.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 svint8_t test_svqsub_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
   return SVE_ACLE_FUNC(svqsub,_s8,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsub_s16_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.u.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z17test_svqsub_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.u.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
 //
 svint16_t test_svqsub_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
   return SVE_ACLE_FUNC(svqsub,_s16,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsub_s32_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.u.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z17test_svqsub_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.u.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
 //
 svint32_t test_svqsub_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
   return SVE_ACLE_FUNC(svqsub,_s32,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsub_s64_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.u.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z17test_svqsub_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.u.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
 //
 svint64_t test_svqsub_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
   return SVE_ACLE_FUNC(svqsub,_s64,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsub_u8_x(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.u.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z16test_svqsub_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.u.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 svuint8_t test_svqsub_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
   return SVE_ACLE_FUNC(svqsub,_u8,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsub_u16_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.u.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z17test_svqsub_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.u.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
 //
 svuint16_t test_svqsub_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
   return SVE_ACLE_FUNC(svqsub,_u16,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsub_u32_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.u.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z17test_svqsub_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.u.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
 //
 svuint32_t test_svqsub_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
   return SVE_ACLE_FUNC(svqsub,_u32,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsub_u64_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.u.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z17test_svqsub_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.u.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
 //
 svuint64_t test_svqsub_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
   return SVE_ACLE_FUNC(svqsub,_u64,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsub_n_s8_z(

 {
   return SVE_ACLE_FUNC(svqsub,_n_u64,_m,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsub_n_s8_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 16 x i8> poison, i8 [[OP2:%.*]], i64 0
 // CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 16 x i8> [[DOTSPLATINSERT]], <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[DOTSPLAT]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.u.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[DOTSPLAT]])
 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svqsub_n_s8_xu10__SVBool_tu10__SVInt8_ta(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 16 x i8> poison, i8 [[OP2:%.*]], i64 0
 // CPP-CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 16 x i8> [[DOTSPLATINSERT]], <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[DOTSPLAT]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.u.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[DOTSPLAT]])
 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 svint8_t test_svqsub_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
   return SVE_ACLE_FUNC(svqsub,_n_s8,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsub_n_s16_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 8 x i16> poison, i16 [[OP2:%.*]], i64 0
 // CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 8 x i16> [[DOTSPLATINSERT]], <vscale x 8 x i16> poison, <vscale x 8 x i32> zeroinitializer
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[DOTSPLAT]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.u.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[DOTSPLAT]])
 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z19test_svqsub_n_s16_xu10__SVBool_tu11__SVInt16_ts(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 8 x i16> poison, i16 [[OP2:%.*]], i64 0
 // CPP-CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 8 x i16> [[DOTSPLATINSERT]], <vscale x 8 x i16> poison, <vscale x 8 x i32> zeroinitializer
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[DOTSPLAT]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.u.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[DOTSPLAT]])
 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
 //
 svint16_t test_svqsub_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
   return SVE_ACLE_FUNC(svqsub,_n_s16,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsub_n_s32_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 4 x i32> poison, i32 [[OP2:%.*]], i64 0
 // CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 4 x i32> [[DOTSPLATINSERT]], <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[DOTSPLAT]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.u.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[DOTSPLAT]])
 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z19test_svqsub_n_s32_xu10__SVBool_tu11__SVInt32_ti(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 4 x i32> poison, i32 [[OP2:%.*]], i64 0
 // CPP-CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 4 x i32> [[DOTSPLATINSERT]], <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[DOTSPLAT]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.u.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[DOTSPLAT]])
 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
 //
 svint32_t test_svqsub_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
   return SVE_ACLE_FUNC(svqsub,_n_s32,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsub_n_s64_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 2 x i64> poison, i64 [[OP2:%.*]], i64 0
 // CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 2 x i64> [[DOTSPLATINSERT]], <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[DOTSPLAT]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.u.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[DOTSPLAT]])
 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z19test_svqsub_n_s64_xu10__SVBool_tu11__SVInt64_tl(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 2 x i64> poison, i64 [[OP2:%.*]], i64 0
 // CPP-CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 2 x i64> [[DOTSPLATINSERT]], <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[DOTSPLAT]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.u.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[DOTSPLAT]])
 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
 //
 svint64_t test_svqsub_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
   return SVE_ACLE_FUNC(svqsub,_n_s64,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsub_n_u8_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 16 x i8> poison, i8 [[OP2:%.*]], i64 0
 // CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 16 x i8> [[DOTSPLATINSERT]], <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[DOTSPLAT]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.u.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[DOTSPLAT]])
 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svqsub_n_u8_xu10__SVBool_tu11__SVUint8_th(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 16 x i8> poison, i8 [[OP2:%.*]], i64 0
 // CPP-CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 16 x i8> [[DOTSPLATINSERT]], <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[DOTSPLAT]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.u.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[DOTSPLAT]])
 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 svuint8_t test_svqsub_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
   return SVE_ACLE_FUNC(svqsub,_n_u8,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsub_n_u16_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 8 x i16> poison, i16 [[OP2:%.*]], i64 0
 // CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 8 x i16> [[DOTSPLATINSERT]], <vscale x 8 x i16> poison, <vscale x 8 x i32> zeroinitializer
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[DOTSPLAT]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.u.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[DOTSPLAT]])
 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z19test_svqsub_n_u16_xu10__SVBool_tu12__SVUint16_tt(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 8 x i16> poison, i16 [[OP2:%.*]], i64 0
 // CPP-CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 8 x i16> [[DOTSPLATINSERT]], <vscale x 8 x i16> poison, <vscale x 8 x i32> zeroinitializer
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[DOTSPLAT]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.u.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[DOTSPLAT]])
 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
 //
 svuint16_t test_svqsub_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
   return SVE_ACLE_FUNC(svqsub,_n_u16,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsub_n_u32_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 4 x i32> poison, i32 [[OP2:%.*]], i64 0
 // CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 4 x i32> [[DOTSPLATINSERT]], <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[DOTSPLAT]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.u.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[DOTSPLAT]])
 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z19test_svqsub_n_u32_xu10__SVBool_tu12__SVUint32_tj(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 4 x i32> poison, i32 [[OP2:%.*]], i64 0
 // CPP-CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 4 x i32> [[DOTSPLATINSERT]], <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[DOTSPLAT]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.u.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[DOTSPLAT]])
 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
 //
 svuint32_t test_svqsub_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
   return SVE_ACLE_FUNC(svqsub,_n_u32,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsub_n_u64_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 2 x i64> poison, i64 [[OP2:%.*]], i64 0
 // CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 2 x i64> [[DOTSPLATINSERT]], <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[DOTSPLAT]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.u.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[DOTSPLAT]])
 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z19test_svqsub_n_u64_xu10__SVBool_tu12__SVUint64_tm(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 2 x i64> poison, i64 [[OP2:%.*]], i64 0
 // CPP-CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 2 x i64> [[DOTSPLATINSERT]], <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[DOTSPLAT]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.u.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[DOTSPLAT]])
 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
 //
 svuint64_t test_svqsub_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
   return SVE_ACLE_FUNC(svqsub,_n_u64,_x,)(pg, op1, op2);
 }

clang/test/CodeGen/aarch64-sve2-intrinsics/acle_sve2_qsubr.c

 //
 svuint64_t test_svqsubr_u64_m(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
   return SVE_ACLE_FUNC(svqsubr,_u64,_m,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsubr_s8_x(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.u.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP1:%.*]])
 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z17test_svqsubr_s8_xu10__SVBool_tu10__SVInt8_tu10__SVInt8_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.u.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP1:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 svint8_t test_svqsubr_s8_x(svbool_t pg, svint8_t op1, svint8_t op2)
 {
   return SVE_ACLE_FUNC(svqsubr,_s8,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsubr_s16_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.u.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP1:%.*]])
 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svqsubr_s16_xu10__SVBool_tu11__SVInt16_tu11__SVInt16_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.u.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP1:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
 //
 svint16_t test_svqsubr_s16_x(svbool_t pg, svint16_t op1, svint16_t op2)
 {
   return SVE_ACLE_FUNC(svqsubr,_s16,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsubr_s32_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.u.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP1:%.*]])
 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svqsubr_s32_xu10__SVBool_tu11__SVInt32_tu11__SVInt32_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.u.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP1:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
 //
 svint32_t test_svqsubr_s32_x(svbool_t pg, svint32_t op1, svint32_t op2)
 {
   return SVE_ACLE_FUNC(svqsubr,_s32,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsubr_s64_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.u.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP1:%.*]])
 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svqsubr_s64_xu10__SVBool_tu11__SVInt64_tu11__SVInt64_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.u.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP1:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
 //
 svint64_t test_svqsubr_s64_x(svbool_t pg, svint64_t op1, svint64_t op2)
 {
   return SVE_ACLE_FUNC(svqsubr,_s64,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsubr_u8_x(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.u.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP1:%.*]])
 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z17test_svqsubr_u8_xu10__SVBool_tu11__SVUint8_tu11__SVUint8_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.u.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP2:%.*]], <vscale x 16 x i8> [[OP1:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 svuint8_t test_svqsubr_u8_x(svbool_t pg, svuint8_t op1, svuint8_t op2)
 {
   return SVE_ACLE_FUNC(svqsubr,_u8,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsubr_u16_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.u.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP1:%.*]])
 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svqsubr_u16_xu10__SVBool_tu12__SVUint16_tu12__SVUint16_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.u.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP2:%.*]], <vscale x 8 x i16> [[OP1:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
 //
 svuint16_t test_svqsubr_u16_x(svbool_t pg, svuint16_t op1, svuint16_t op2)
 {
   return SVE_ACLE_FUNC(svqsubr,_u16,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsubr_u32_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.u.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP1:%.*]])
 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svqsubr_u32_xu10__SVBool_tu12__SVUint32_tu12__SVUint32_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.u.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP2:%.*]], <vscale x 4 x i32> [[OP1:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
 //
 svuint32_t test_svqsubr_u32_x(svbool_t pg, svuint32_t op1, svuint32_t op2)
 {
   return SVE_ACLE_FUNC(svqsubr,_u32,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsubr_u64_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.u.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP1:%.*]])
 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z18test_svqsubr_u64_xu10__SVBool_tu12__SVUint64_tu12__SVUint64_t(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.u.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP2:%.*]], <vscale x 2 x i64> [[OP1:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
 //
 svuint64_t test_svqsubr_u64_x(svbool_t pg, svuint64_t op1, svuint64_t op2)
 {
   return SVE_ACLE_FUNC(svqsubr,_u64,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsubr_n_s8_z(

 {
   return SVE_ACLE_FUNC(svqsubr,_n_u64,_m,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsubr_n_s8_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 16 x i8> poison, i8 [[OP2:%.*]], i64 0
 // CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 16 x i8> [[DOTSPLATINSERT]], <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[DOTSPLAT]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.u.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[DOTSPLAT]], <vscale x 16 x i8> [[OP1:%.*]])
 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z19test_svqsubr_n_s8_xu10__SVBool_tu10__SVInt8_ta(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 16 x i8> poison, i8 [[OP2:%.*]], i64 0
 // CPP-CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 16 x i8> [[DOTSPLATINSERT]], <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[DOTSPLAT]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.u.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[DOTSPLAT]], <vscale x 16 x i8> [[OP1:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 svint8_t test_svqsubr_n_s8_x(svbool_t pg, svint8_t op1, int8_t op2)
 {
   return SVE_ACLE_FUNC(svqsubr,_n_s8,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsubr_n_s16_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 8 x i16> poison, i16 [[OP2:%.*]], i64 0
 // CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 8 x i16> [[DOTSPLATINSERT]], <vscale x 8 x i16> poison, <vscale x 8 x i32> zeroinitializer
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[DOTSPLAT]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.u.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[DOTSPLAT]], <vscale x 8 x i16> [[OP1:%.*]])
 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z20test_svqsubr_n_s16_xu10__SVBool_tu11__SVInt16_ts(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 8 x i16> poison, i16 [[OP2:%.*]], i64 0
 // CPP-CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 8 x i16> [[DOTSPLATINSERT]], <vscale x 8 x i16> poison, <vscale x 8 x i32> zeroinitializer
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[DOTSPLAT]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.u.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[DOTSPLAT]], <vscale x 8 x i16> [[OP1:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
 //
 svint16_t test_svqsubr_n_s16_x(svbool_t pg, svint16_t op1, int16_t op2)
 {
   return SVE_ACLE_FUNC(svqsubr,_n_s16,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsubr_n_s32_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 4 x i32> poison, i32 [[OP2:%.*]], i64 0
 // CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 4 x i32> [[DOTSPLATINSERT]], <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[DOTSPLAT]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.u.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[DOTSPLAT]], <vscale x 4 x i32> [[OP1:%.*]])
 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z20test_svqsubr_n_s32_xu10__SVBool_tu11__SVInt32_ti(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 4 x i32> poison, i32 [[OP2:%.*]], i64 0
 // CPP-CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 4 x i32> [[DOTSPLATINSERT]], <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[DOTSPLAT]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.u.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[DOTSPLAT]], <vscale x 4 x i32> [[OP1:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
 //
 svint32_t test_svqsubr_n_s32_x(svbool_t pg, svint32_t op1, int32_t op2)
 {
   return SVE_ACLE_FUNC(svqsubr,_n_s32,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsubr_n_s64_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 2 x i64> poison, i64 [[OP2:%.*]], i64 0
 // CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 2 x i64> [[DOTSPLATINSERT]], <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[DOTSPLAT]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.u.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[DOTSPLAT]], <vscale x 2 x i64> [[OP1:%.*]])
 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z20test_svqsubr_n_s64_xu10__SVBool_tu11__SVInt64_tl(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 2 x i64> poison, i64 [[OP2:%.*]], i64 0
 // CPP-CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 2 x i64> [[DOTSPLATINSERT]], <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[DOTSPLAT]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.u.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[DOTSPLAT]], <vscale x 2 x i64> [[OP1:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
 //
 svint64_t test_svqsubr_n_s64_x(svbool_t pg, svint64_t op1, int64_t op2)
 {
   return SVE_ACLE_FUNC(svqsubr,_n_s64,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsubr_n_u8_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 16 x i8> poison, i8 [[OP2:%.*]], i64 0
 // CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 16 x i8> [[DOTSPLATINSERT]], <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer
-// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[DOTSPLAT]])
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.u.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[DOTSPLAT]], <vscale x 16 x i8> [[OP1:%.*]])
 // CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 // CPP-CHECK-LABEL: @_Z19test_svqsubr_n_u8_xu10__SVBool_tu11__SVUint8_th(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 16 x i8> poison, i8 [[OP2:%.*]], i64 0
 // CPP-CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 16 x i8> [[DOTSPLATINSERT]], <vscale x 16 x i8> poison, <vscale x 16 x i32> zeroinitializer
-// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uqsubr.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[DOTSPLAT]])
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.u.nxv16i8(<vscale x 16 x i1> [[PG:%.*]], <vscale x 16 x i8> [[DOTSPLAT]], <vscale x 16 x i8> [[OP1:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 svuint8_t test_svqsubr_n_u8_x(svbool_t pg, svuint8_t op1, uint8_t op2)
 {
   return SVE_ACLE_FUNC(svqsubr,_n_u8,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsubr_n_u16_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 8 x i16> poison, i16 [[OP2:%.*]], i64 0
 // CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 8 x i16> [[DOTSPLATINSERT]], <vscale x 8 x i16> poison, <vscale x 8 x i32> zeroinitializer
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[DOTSPLAT]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.u.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[DOTSPLAT]], <vscale x 8 x i16> [[OP1:%.*]])
 // CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z20test_svqsubr_n_u16_xu10__SVBool_tu12__SVUint16_tt(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 8 x i16> poison, i16 [[OP2:%.*]], i64 0
 // CPP-CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 8 x i16> [[DOTSPLATINSERT]], <vscale x 8 x i16> poison, <vscale x 8 x i32> zeroinitializer
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqsubr.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[DOTSPLAT]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.u.nxv8i16(<vscale x 8 x i1> [[TMP0]], <vscale x 8 x i16> [[DOTSPLAT]], <vscale x 8 x i16> [[OP1:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 8 x i16> [[TMP1]]
 //
 svuint16_t test_svqsubr_n_u16_x(svbool_t pg, svuint16_t op1, uint16_t op2)
 {
   return SVE_ACLE_FUNC(svqsubr,_n_u16,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsubr_n_u32_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 4 x i32> poison, i32 [[OP2:%.*]], i64 0
 // CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 4 x i32> [[DOTSPLATINSERT]], <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[DOTSPLAT]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.u.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[DOTSPLAT]], <vscale x 4 x i32> [[OP1:%.*]])
 // CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z20test_svqsubr_n_u32_xu10__SVBool_tu12__SVUint32_tj(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 4 x i32> poison, i32 [[OP2:%.*]], i64 0
 // CPP-CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 4 x i32> [[DOTSPLATINSERT]], <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.uqsubr.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[DOTSPLAT]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.u.nxv4i32(<vscale x 4 x i1> [[TMP0]], <vscale x 4 x i32> [[DOTSPLAT]], <vscale x 4 x i32> [[OP1:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 4 x i32> [[TMP1]]
 //
 svuint32_t test_svqsubr_n_u32_x(svbool_t pg, svuint32_t op1, uint32_t op2)
 {
   return SVE_ACLE_FUNC(svqsubr,_n_u32,_x,)(pg, op1, op2);
 }
 
 // CHECK-LABEL: @test_svqsubr_n_u64_x(
 // CHECK-NEXT:  entry:
 // CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 2 x i64> poison, i64 [[OP2:%.*]], i64 0
 // CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 2 x i64> [[DOTSPLATINSERT]], <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer
-// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[DOTSPLAT]])
+// CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.u.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[DOTSPLAT]], <vscale x 2 x i64> [[OP1:%.*]])
 // CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z20test_svqsubr_n_u64_xu10__SVBool_tu12__SVUint64_tm(
 // CPP-CHECK-NEXT:  entry:
 // CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call <vscale x 2 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv2i1(<vscale x 16 x i1> [[PG:%.*]])
 // CPP-CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 2 x i64> poison, i64 [[OP2:%.*]], i64 0
 // CPP-CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <vscale x 2 x i64> [[DOTSPLATINSERT]], <vscale x 2 x i64> poison, <vscale x 2 x i32> zeroinitializer
-// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.uqsubr.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[DOTSPLAT]])
+// CPP-CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.u.nxv2i64(<vscale x 2 x i1> [[TMP0]], <vscale x 2 x i64> [[DOTSPLAT]], <vscale x 2 x i64> [[OP1:%.*]])
 // CPP-CHECK-NEXT:    ret <vscale x 2 x i64> [[TMP1]]
 //
 svuint64_t test_svqsubr_n_u64_x(svbool_t pg, svuint64_t op1, uint64_t op2)
 {
   return SVE_ACLE_FUNC(svqsubr,_n_u64,_x,)(pg, op1, op2);
 }

llvm/include/llvm/IR/IntrinsicsAArch64.td

 def int_aarch64_sve_sqrdmlsh      : AdvSIMD_3VectorArg_Intrinsic;
 def int_aarch64_sve_sqrdmlsh_lane : AdvSIMD_3VectorArgIndexed_Intrinsic;
 def int_aarch64_sve_sqrdmulh      : AdvSIMD_2VectorArg_Intrinsic;
 def int_aarch64_sve_sqrdmulh_lane : AdvSIMD_2VectorArgIndexed_Intrinsic;
 def int_aarch64_sve_sqrshl        : AdvSIMD_Pred2VectorArg_Intrinsic;
 def int_aarch64_sve_sqshl         : AdvSIMD_Pred2VectorArg_Intrinsic;
 def int_aarch64_sve_sqshlu        : AdvSIMD_SVE_ShiftByImm_Intrinsic;
 def int_aarch64_sve_sqsub         : AdvSIMD_Pred2VectorArg_Intrinsic;
+def int_aarch64_sve_sqsub_u       : AdvSIMD_Pred2VectorArg_Intrinsic;
 def int_aarch64_sve_sqsubr        : AdvSIMD_Pred2VectorArg_Intrinsic;
 def int_aarch64_sve_srhadd        : AdvSIMD_Pred2VectorArg_Intrinsic;
 def int_aarch64_sve_sri           : AdvSIMD_2VectorArgIndexed_Intrinsic;
 def int_aarch64_sve_srshl         : AdvSIMD_Pred2VectorArg_Intrinsic;
 def int_aarch64_sve_srshr         : AdvSIMD_SVE_ShiftByImm_Intrinsic;
 def int_aarch64_sve_srsra         : AdvSIMD_2VectorArgIndexed_Intrinsic;
 def int_aarch64_sve_ssra          : AdvSIMD_2VectorArgIndexed_Intrinsic;
 def int_aarch64_sve_suqadd        : AdvSIMD_Pred2VectorArg_Intrinsic;
 def int_aarch64_sve_uaba          : AdvSIMD_3VectorArg_Intrinsic;
 def int_aarch64_sve_uhadd         : AdvSIMD_Pred2VectorArg_Intrinsic;
 def int_aarch64_sve_uhsub         : AdvSIMD_Pred2VectorArg_Intrinsic;
 def int_aarch64_sve_uhsubr        : AdvSIMD_Pred2VectorArg_Intrinsic;
 def int_aarch64_sve_uqadd         : AdvSIMD_Pred2VectorArg_Intrinsic;
 def int_aarch64_sve_uqrshl        : AdvSIMD_Pred2VectorArg_Intrinsic;
 def int_aarch64_sve_uqshl         : AdvSIMD_Pred2VectorArg_Intrinsic;
 def int_aarch64_sve_uqsub         : AdvSIMD_Pred2VectorArg_Intrinsic;
+def int_aarch64_sve_uqsub_u       : AdvSIMD_Pred2VectorArg_Intrinsic;
 def int_aarch64_sve_uqsubr        : AdvSIMD_Pred2VectorArg_Intrinsic;
 def int_aarch64_sve_urecpe        : AdvSIMD_Merged1VectorArg_Intrinsic;
 def int_aarch64_sve_urhadd        : AdvSIMD_Pred2VectorArg_Intrinsic;
 def int_aarch64_sve_urshl         : AdvSIMD_Pred2VectorArg_Intrinsic;
 def int_aarch64_sve_urshr         : AdvSIMD_SVE_ShiftByImm_Intrinsic;
 def int_aarch64_sve_ursqrte       : AdvSIMD_Merged1VectorArg_Intrinsic;
 def int_aarch64_sve_ursra         : AdvSIMD_2VectorArgIndexed_Intrinsic;
 def int_aarch64_sve_usqadd        : AdvSIMD_Pred2VectorArg_Intrinsic;

llvm/lib/Target/AArch64/AArch64ISelLowering.cpp

return DAG.getNode(AArch64ISD::SDIV_PRED, SDLoc(N), N->getValueType(0),
                        N->getOperand(1), N->getOperand(2), N->getOperand(3));
   case Intrinsic::aarch64_sve_udiv_u:
     return DAG.getNode(AArch64ISD::UDIV_PRED, SDLoc(N), N->getValueType(0),
                        N->getOperand(1), N->getOperand(2), N->getOperand(3));
   case Intrinsic::aarch64_sve_sqadd:
     return convertMergedOpToPredOp(N, ISD::SADDSAT, DAG, true);
   case Intrinsic::aarch64_sve_sqsub:
     return convertMergedOpToPredOp(N, ISD::SSUBSAT, DAG, true);
+  case Intrinsic::aarch64_sve_sqsub_u:
+    return DAG.getNode(ISD::SSUBSAT, SDLoc(N), N->getValueType(0),
+                       N->getOperand(2), N->getOperand(3));
   case Intrinsic::aarch64_sve_uqadd:
     return convertMergedOpToPredOp(N, ISD::UADDSAT, DAG, true);
   case Intrinsic::aarch64_sve_uqsub:
     return convertMergedOpToPredOp(N, ISD::USUBSAT, DAG, true);
+  case Intrinsic::aarch64_sve_uqsub_u:
+    return DAG.getNode(ISD::USUBSAT, SDLoc(N), N->getValueType(0),
+                       N->getOperand(2), N->getOperand(3));
   case Intrinsic::aarch64_sve_sqadd_x:
     return DAG.getNode(ISD::SADDSAT, SDLoc(N), N->getValueType(0),
                        N->getOperand(1), N->getOperand(2));
   case Intrinsic::aarch64_sve_sqsub_x:
     return DAG.getNode(ISD::SSUBSAT, SDLoc(N), N->getValueType(0),
                        N->getOperand(1), N->getOperand(2));
   case Intrinsic::aarch64_sve_uqadd_x:
     return DAG.getNode(ISD::UADDSAT, SDLoc(N), N->getValueType(0),

llvm/test/CodeGen/AArch64/sve2-intrinsics-uniform-dsp-undef.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve2 < %s | FileCheck %s
+
+;
+; SQSUB
+;
+
+define <vscale x 16 x i8> @sqsub_i8_u(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b) {
+; CHECK-LABEL: sqsub_i8_u:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sqsub z0.b, z0.b, z1.b
+; CHECK-NEXT:    ret
+  %out = call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.u.nxv16i8(<vscale x 16 x i1> %pg,
+                                                                 <vscale x 16 x i8> %a,
+                                                                 <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %out
+}
+
+define <vscale x 8 x i16> @sqsub_i16_u(<vscale x 8 x i1> %pg, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b) {
+; CHECK-LABEL: sqsub_i16_u:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sqsub z0.h, z0.h, z1.h
+; CHECK-NEXT:    ret
+  %out = call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.u.nxv8i16(<vscale x 8 x i1> %pg,
+                                                                 <vscale x 8 x i16> %a,
+                                                                 <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %out
+}
+
+define <vscale x 4 x i32> @sqsub_i32_u(<vscale x 4 x i1> %pg, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b) {
+; CHECK-LABEL: sqsub_i32_u:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sqsub z0.s, z0.s, z1.s
+; CHECK-NEXT:    ret
+  %out = call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.u.nxv4i32(<vscale x 4 x i1> %pg,
+                                                                 <vscale x 4 x i32> %a,
+                                                                 <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %out
+}
+
+define <vscale x 2 x i64> @sqsub_i64_u(<vscale x 2 x i1> %pg, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b) {
+; CHECK-LABEL: sqsub_i64_u:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sqsub z0.d, z0.d, z1.d
+; CHECK-NEXT:    ret
+  %out = call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.u.nxv2i64(<vscale x 2 x i1> %pg,
+                                                                 <vscale x 2 x i64> %a,
+                                                                 <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %out
+}
+
+;
+; UQSUB
+;
+
+define <vscale x 16 x i8> @uqsub_i8_u(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b) {
+; CHECK-LABEL: uqsub_i8_u:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    uqsub z0.b, z0.b, z1.b
+; CHECK-NEXT:    ret
+  %out = call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.u.nxv16i8(<vscale x 16 x i1> %pg,
+                                                                 <vscale x 16 x i8> %a,
+                                                                 <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %out
+}
+
+define <vscale x 8 x i16> @uqsub_i16_u(<vscale x 8 x i1> %pg, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b) {
+; CHECK-LABEL: uqsub_i16_u:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    uqsub z0.h, z0.h, z1.h
+; CHECK-NEXT:    ret
+  %out = call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.u.nxv8i16(<vscale x 8 x i1> %pg,
+                                                                 <vscale x 8 x i16> %a,
+                                                                 <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %out
+}
+
+define <vscale x 4 x i32> @uqsub_i32_u(<vscale x 4 x i1> %pg, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b) {
+; CHECK-LABEL: uqsub_i32_u:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    uqsub z0.s, z0.s, z1.s
+; CHECK-NEXT:    ret
+  %out = call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.u.nxv4i32(<vscale x 4 x i1> %pg,
+                                                                 <vscale x 4 x i32> %a,
+                                                                 <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %out
+}
+
+define <vscale x 2 x i64> @uqsub_i64_u(<vscale x 2 x i1> %pg, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b) {
+; CHECK-LABEL: uqsub_i64_u:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    uqsub z0.d, z0.d, z1.d
+; CHECK-NEXT:    ret
+  %out = call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.u.nxv2i64(<vscale x 2 x i1> %pg,
+                                                                 <vscale x 2 x i64> %a,
+                                                                 <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %out
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+declare <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+declare <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+declare <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+declare <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+declare <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+declare <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)