From ce70e2899879e092b153a4078b993833b6696713 Mon Sep 17 00:00:00 2001
From: Danilo Carvalho Grael <danilo.carvalho.grael@huawei.com>
Date: Thu, 20 Feb 2020 11:24:50 -0500
Subject: [PATCH] [AArch64][SVE] Add intrinsics for SVE2 bitwise ternary
 operations

Summary:
Add intrinsics for the following operations:
- eor3, bcax
- bsl, bsl1n, bsl2n, nbsl

Reviewers: kmclaughlin, c-rhodes, sdesmalen, efriedma, rengolin

Reviewed By: efriedma

Subscribers: tschuett, kristof.beyls, hiraditya, rkruppe, psnobl, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D74785
---
 llvm/include/llvm/IR/IntrinsicsAArch64.td     |   9 +-
 .../lib/Target/AArch64/AArch64SVEInstrInfo.td |  12 +-
 llvm/lib/Target/AArch64/SVEInstrFormats.td    |   7 +-
 .../CodeGen/AArch64/sve2-bitwise-ternary.ll   | 284 ++++++++++++++++++
 4 files changed, 304 insertions(+), 8 deletions(-)
 create mode 100644 llvm/test/CodeGen/AArch64/sve2-bitwise-ternary.ll

diff --git a/llvm/include/llvm/IR/IntrinsicsAArch64.td b/llvm/include/llvm/IR/IntrinsicsAArch64.td
index a3fb5fe840a4..a3234b3bdd5a 100644
--- a/llvm/include/llvm/IR/IntrinsicsAArch64.td
+++ b/llvm/include/llvm/IR/IntrinsicsAArch64.td
@@ -2001,10 +2001,17 @@ def int_aarch64_sve_sbclt       : AdvSIMD_3VectorArg_Intrinsic;
 //
 // SVE2 - Polynomial arithmetic
 //
-
 def int_aarch64_sve_eorbt       : AdvSIMD_3VectorArg_Intrinsic;
 def int_aarch64_sve_eortb       : AdvSIMD_3VectorArg_Intrinsic;
 def int_aarch64_sve_pmullb_pair : AdvSIMD_2VectorArg_Intrinsic;
 def int_aarch64_sve_pmullt_pair : AdvSIMD_2VectorArg_Intrinsic;
 
+// SVE2 bitwise ternary operations.
+def int_aarch64_sve_eor3   : AdvSIMD_3VectorArg_Intrinsic;
+def int_aarch64_sve_bcax   : AdvSIMD_3VectorArg_Intrinsic;
+def int_aarch64_sve_bsl    : AdvSIMD_3VectorArg_Intrinsic;
+def int_aarch64_sve_bsl1n  : AdvSIMD_3VectorArg_Intrinsic;
+def int_aarch64_sve_bsl2n  : AdvSIMD_3VectorArg_Intrinsic;
+def int_aarch64_sve_nbsl   : AdvSIMD_3VectorArg_Intrinsic;
+
 }
diff --git a/llvm/lib/Target/AArch64/AArch64SVEInstrInfo.td b/llvm/lib/Target/AArch64/AArch64SVEInstrInfo.td
index f8819c6b9b14..59c478a3a386 100644
--- a/llvm/lib/Target/AArch64/AArch64SVEInstrInfo.td
+++ b/llvm/lib/Target/AArch64/AArch64SVEInstrInfo.td
@@ -1771,12 +1771,12 @@ let Predicates = [HasSVE2] in {
   defm FMLSLT_ZZZ_SHH : sve2_fp_mla_long<0b11, "fmlslt", int_aarch64_sve_fmlslt>;
 
   // SVE2 bitwise ternary operations
-  defm EOR3_ZZZZ_D  : sve2_int_bitwise_ternary_op<0b000, "eor3">;
-  defm BCAX_ZZZZ_D  : sve2_int_bitwise_ternary_op<0b010, "bcax">;
-  def BSL_ZZZZ_D    : sve2_int_bitwise_ternary_op_d<0b001, "bsl">;
-  def BSL1N_ZZZZ_D  : sve2_int_bitwise_ternary_op_d<0b011, "bsl1n">;
-  def BSL2N_ZZZZ_D  : sve2_int_bitwise_ternary_op_d<0b101, "bsl2n">;
-  def NBSL_ZZZZ_D   : sve2_int_bitwise_ternary_op_d<0b111, "nbsl">;
+  defm EOR3_ZZZZ  : sve2_int_bitwise_ternary_op<0b000, "eor3",  int_aarch64_sve_eor3>;
+  defm BCAX_ZZZZ  : sve2_int_bitwise_ternary_op<0b010, "bcax",  int_aarch64_sve_bcax>;
+  defm BSL_ZZZZ   : sve2_int_bitwise_ternary_op<0b001, "bsl",   int_aarch64_sve_bsl>;
+  defm BSL1N_ZZZZ : sve2_int_bitwise_ternary_op<0b011, "bsl1n", int_aarch64_sve_bsl1n>;
+  defm BSL2N_ZZZZ : sve2_int_bitwise_ternary_op<0b101, "bsl2n", int_aarch64_sve_bsl2n>;
+  defm NBSL_ZZZZ  : sve2_int_bitwise_ternary_op<0b111, "nbsl",  int_aarch64_sve_nbsl>;
 
   // SVE2 bitwise xor and rotate right by immediate
   defm XAR_ZZZI : sve2_int_rotate_right_imm<"xar">;
diff --git a/llvm/lib/Target/AArch64/SVEInstrFormats.td b/llvm/lib/Target/AArch64/SVEInstrFormats.td
index 757743084721..8c02b3a95dfe 100644
--- a/llvm/lib/Target/AArch64/SVEInstrFormats.td
+++ b/llvm/lib/Target/AArch64/SVEInstrFormats.td
@@ -3766,7 +3766,7 @@ class sve2_int_bitwise_ternary_op_d<bits<3> opc, string asm>
   let ElementSize = ElementSizeNone;
 }
 
-multiclass sve2_int_bitwise_ternary_op<bits<3> opc, string asm> {
+multiclass sve2_int_bitwise_ternary_op<bits<3> opc, string asm, SDPatternOperator op> {
   def NAME : sve2_int_bitwise_ternary_op_d<opc, asm>;
 
   def : InstAlias<asm # "\t$Zdn, $Zdn, $Zm, $Zk",
@@ -3775,6 +3775,11 @@ multiclass sve2_int_bitwise_ternary_op<bits<3> opc, string asm> {
                   (!cast<Instruction>(NAME) ZPR16:$Zdn, ZPR16:$Zm, ZPR16:$Zk), 1>;
   def : InstAlias<asm # "\t$Zdn, $Zdn, $Zm, $Zk",
                   (!cast<Instruction>(NAME) ZPR32:$Zdn, ZPR32:$Zm, ZPR32:$Zk), 1>;
+
+  def : SVE_3_Op_Pat<nxv16i8, op, nxv16i8, nxv16i8, nxv16i8, !cast<Instruction>(NAME)>;
+  def : SVE_3_Op_Pat<nxv8i16, op, nxv8i16, nxv8i16, nxv8i16, !cast<Instruction>(NAME)>;
+  def : SVE_3_Op_Pat<nxv4i32, op, nxv4i32, nxv4i32, nxv4i32, !cast<Instruction>(NAME)>;
+  def : SVE_3_Op_Pat<nxv2i64, op, nxv2i64, nxv2i64, nxv2i64, !cast<Instruction>(NAME)>;
 }
 
 class sve2_int_rotate_right_imm<bits<4> tsz8_64, string asm,
diff --git a/llvm/test/CodeGen/AArch64/sve2-bitwise-ternary.ll b/llvm/test/CodeGen/AArch64/sve2-bitwise-ternary.ll
new file mode 100644
index 000000000000..8745f7f96e64
--- /dev/null
+++ b/llvm/test/CodeGen/AArch64/sve2-bitwise-ternary.ll
@@ -0,0 +1,284 @@
+; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve2 < %s | FileCheck %s
+
+;
+; EOR3 (vector, bitwise, unpredicated)
+;
+define <vscale x 16 x i8> @eor3_i8(<vscale x 16 x i8> %a,
+                                   <vscale x 16 x i8> %b,
+                                   <vscale x 16 x i8> %c) {
+; CHECK-LABEL: eor3_i8
+; CHECK: eor3 z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 16 x i8> @llvm.aarch64.sve.eor3.nxv16i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b, <vscale x 16 x i8> %c)
+  ret <vscale x 16 x i8> %res
+}
+
+define <vscale x 8 x i16> @eor3_i16(<vscale x 8 x i16> %a,
+                                    <vscale x 8 x i16> %b,
+                                    <vscale x 8 x i16> %c) {
+; CHECK-LABEL: eor3_i16
+; CHECK: eor3 z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 8 x i16> @llvm.aarch64.sve.eor3.nxv8i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b, <vscale x 8 x i16> %c)
+  ret <vscale x 8 x i16> %res
+}
+
+define <vscale x 4 x i32> @eor3_i32(<vscale x 4 x i32> %a,
+                                    <vscale x 4 x i32> %b,
+                                    <vscale x 4 x i32> %c) {
+; CHECK-LABEL: eor3_i32
+; CHECK: eor3 z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 4 x i32> @llvm.aarch64.sve.eor3.nxv4i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b, <vscale x 4 x i32> %c)
+  ret <vscale x 4 x i32> %res
+}
+
+define <vscale x 2 x i64> @eor3_i64(<vscale x 2 x i64> %a,
+                                    <vscale x 2 x i64> %b,
+                                    <vscale x 2 x i64> %c) {
+; CHECK-LABEL: eor3_i64
+; CHECK: eor3 z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 2 x i64> @llvm.aarch64.sve.eor3.nxv2i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b, <vscale x 2 x i64> %c)
+  ret <vscale x 2 x i64> %res
+}
+
+;
+; BCAX (vector, bitwise, unpredicated)
+;
+define <vscale x 16 x i8> @bcax_i8(<vscale x 16 x i8> %a,
+                                   <vscale x 16 x i8> %b,
+                                   <vscale x 16 x i8> %c) {
+; CHECK-LABEL: bcax_i8
+; CHECK: bcax z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 16 x i8> @llvm.aarch64.sve.bcax.nxv16i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b, <vscale x 16 x i8> %c)
+  ret <vscale x 16 x i8> %res
+}
+
+define <vscale x 8 x i16> @bcax_i16(<vscale x 8 x i16> %a,
+                                    <vscale x 8 x i16> %b,
+                                    <vscale x 8 x i16> %c) {
+; CHECK-LABEL: bcax_i16
+; CHECK: bcax z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 8 x i16> @llvm.aarch64.sve.bcax.nxv8i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b, <vscale x 8 x i16> %c)
+  ret <vscale x 8 x i16> %res
+}
+
+define <vscale x 4 x i32> @bcax_i32(<vscale x 4 x i32> %a,
+                                    <vscale x 4 x i32> %b,
+                                    <vscale x 4 x i32> %c) {
+; CHECK-LABEL: bcax_i32
+; CHECK: bcax z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 4 x i32> @llvm.aarch64.sve.bcax.nxv4i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b, <vscale x 4 x i32> %c)
+  ret <vscale x 4 x i32> %res
+}
+
+define <vscale x 2 x i64> @bcax_i64(<vscale x 2 x i64> %a,
+                                    <vscale x 2 x i64> %b,
+                                    <vscale x 2 x i64> %c) {
+; CHECK-LABEL: bcax_i64
+; CHECK: bcax z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 2 x i64> @llvm.aarch64.sve.bcax.nxv2i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b, <vscale x 2 x i64> %c)
+  ret <vscale x 2 x i64> %res
+}
+
+;
+; BSL (vector, bitwise, unpredicated)
+;
+define <vscale x 16 x i8> @bsl_i8(<vscale x 16 x i8> %a,
+                                  <vscale x 16 x i8> %b,
+                                  <vscale x 16 x i8> %c) {
+; CHECK-LABEL: bsl_i8
+; CHECK: bsl z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl.nxv16i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b, <vscale x 16 x i8> %c)
+  ret <vscale x 16 x i8> %res
+}
+
+define <vscale x 8 x i16> @bsl_i16(<vscale x 8 x i16> %a,
+                                   <vscale x 8 x i16> %b,
+                                   <vscale x 8 x i16> %c) {
+; CHECK-LABEL: bsl_i16
+; CHECK: bsl z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl.nxv8i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b, <vscale x 8 x i16> %c)
+  ret <vscale x 8 x i16> %res
+}
+
+define <vscale x 4 x i32> @bsl_i32(<vscale x 4 x i32> %a,
+                                   <vscale x 4 x i32> %b,
+                                   <vscale x 4 x i32> %c) {
+; CHECK-LABEL: bsl_i32
+; CHECK: bsl z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl.nxv4i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b, <vscale x 4 x i32> %c)
+  ret <vscale x 4 x i32> %res
+}
+
+define <vscale x 2 x i64> @bsl_i64(<vscale x 2 x i64> %a,
+                                   <vscale x 2 x i64> %b,
+                                   <vscale x 2 x i64> %c) {
+; CHECK-LABEL: bsl_i64
+; CHECK: bsl z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl.nxv2i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b, <vscale x 2 x i64> %c)
+  ret <vscale x 2 x i64> %res
+}
+
+;
+; BSL1N (vector, bitwise, unpredicated)
+;
+define <vscale x 16 x i8> @bsl1n_i8(<vscale x 16 x i8> %a,
+                                    <vscale x 16 x i8> %b,
+                                    <vscale x 16 x i8> %c) {
+; CHECK-LABEL: bsl1n_i8
+; CHECK: bsl1n z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl1n.nxv16i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b, <vscale x 16 x i8> %c)
+  ret <vscale x 16 x i8> %res
+}
+
+define <vscale x 8 x i16> @bsl1n_i16(<vscale x 8 x i16> %a,
+                                     <vscale x 8 x i16> %b,
+                                     <vscale x 8 x i16> %c) {
+; CHECK-LABEL: bsl1n_i16
+; CHECK: bsl1n z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl1n.nxv8i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b, <vscale x 8 x i16> %c)
+  ret <vscale x 8 x i16> %res
+}
+
+define <vscale x 4 x i32> @bsl1n_i32(<vscale x 4 x i32> %a,
+                                     <vscale x 4 x i32> %b,
+                                     <vscale x 4 x i32> %c) {
+; CHECK-LABEL: bsl1n_i32
+; CHECK: bsl1n z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl1n.nxv4i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b, <vscale x 4 x i32> %c)
+  ret <vscale x 4 x i32> %res
+}
+
+define <vscale x 2 x i64> @bsl1n_i64(<vscale x 2 x i64> %a,
+                                     <vscale x 2 x i64> %b,
+                                     <vscale x 2 x i64> %c) {
+; CHECK-LABEL: bsl1n_i64
+; CHECK: bsl1n z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl1n.nxv2i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b, <vscale x 2 x i64> %c)
+  ret <vscale x 2 x i64> %res
+}
+
+;
+; BSL2N (vector, bitwise, unpredicated)
+;
+define <vscale x 16 x i8> @bsl2n_i8(<vscale x 16 x i8> %a,
+                                    <vscale x 16 x i8> %b,
+                                    <vscale x 16 x i8> %c) {
+; CHECK-LABEL: bsl2n_i8
+; CHECK: bsl2n z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 16 x i8> @llvm.aarch64.sve.bsl2n.nxv16i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b, <vscale x 16 x i8> %c)
+  ret <vscale x 16 x i8> %res
+}
+
+define <vscale x 8 x i16> @bsl2n_i16(<vscale x 8 x i16> %a,
+                                     <vscale x 8 x i16> %b,
+                                     <vscale x 8 x i16> %c) {
+; CHECK-LABEL: bsl2n_i16
+; CHECK: bsl2n z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 8 x i16> @llvm.aarch64.sve.bsl2n.nxv8i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b, <vscale x 8 x i16> %c)
+  ret <vscale x 8 x i16> %res
+}
+
+define <vscale x 4 x i32> @bsl2n_i32(<vscale x 4 x i32> %a,
+                                     <vscale x 4 x i32> %b,
+                                     <vscale x 4 x i32> %c) {
+; CHECK-LABEL: bsl2n_i32
+; CHECK: bsl2n z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 4 x i32> @llvm.aarch64.sve.bsl2n.nxv4i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b, <vscale x 4 x i32> %c)
+  ret <vscale x 4 x i32> %res
+}
+
+define <vscale x 2 x i64> @bsl2n_i64(<vscale x 2 x i64> %a,
+                                     <vscale x 2 x i64> %b,
+                                     <vscale x 2 x i64> %c) {
+; CHECK-LABEL: bsl2n_i64
+; CHECK: bsl2n z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 2 x i64> @llvm.aarch64.sve.bsl2n.nxv2i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b, <vscale x 2 x i64> %c)
+  ret <vscale x 2 x i64> %res
+}
+
+;
+; NBSL (vector, bitwise, unpredicated)
+;
+define <vscale x 16 x i8> @nbsl_i8(<vscale x 16 x i8> %a,
+                                   <vscale x 16 x i8> %b,
+                                   <vscale x 16 x i8> %c) {
+; CHECK-LABEL: nbsl_i8
+; CHECK: nbsl z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 16 x i8> @llvm.aarch64.sve.nbsl.nxv16i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b, <vscale x 16 x i8> %c)
+  ret <vscale x 16 x i8> %res
+}
+
+define <vscale x 8 x i16> @nbsl_i16(<vscale x 8 x i16> %a,
+                                    <vscale x 8 x i16> %b,
+                                    <vscale x 8 x i16> %c) {
+; CHECK-LABEL: nbsl_i16
+; CHECK: nbsl z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 8 x i16> @llvm.aarch64.sve.nbsl.nxv8i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b, <vscale x 8 x i16> %c)
+  ret <vscale x 8 x i16> %res
+}
+
+define <vscale x 4 x i32> @nbsl_i32(<vscale x 4 x i32> %a,
+                                    <vscale x 4 x i32> %b,
+                                    <vscale x 4 x i32> %c) {
+; CHECK-LABEL: nbsl_i32
+; CHECK: nbsl z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 4 x i32> @llvm.aarch64.sve.nbsl.nxv4i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b, <vscale x 4 x i32> %c)
+  ret <vscale x 4 x i32> %res
+}
+
+define <vscale x 2 x i64> @nbsl_i64(<vscale x 2 x i64> %a,
+                                    <vscale x 2 x i64> %b,
+                                    <vscale x 2 x i64> %c) {
+; CHECK-LABEL: nbsl_i64
+; CHECK: nbsl z0.d, z0.d, z1.d, z2.d
+; CHECK-NEXT: ret
+  %res = call <vscale x 2 x i64> @llvm.aarch64.sve.nbsl.nxv2i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b, <vscale x 2 x i64> %c)
+  ret <vscale x 2 x i64> %res
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.eor3.nxv16i8(<vscale x 16 x i8>,<vscale x 16 x i8>,<vscale x 16 x i8>)
+declare <vscale x 8 x i16> @llvm.aarch64.sve.eor3.nxv8i16(<vscale x 8 x i16>,<vscale x 8 x i16>,<vscale x 8 x i16>)
+declare <vscale x 4 x i32> @llvm.aarch64.sve.eor3.nxv4i32(<vscale x 4 x i32>,<vscale x 4 x i32>,<vscale x 4 x i32>)
+declare <vscale x 2 x i64> @llvm.aarch64.sve.eor3.nxv2i64(<vscale x 2 x i64>,<vscale x 2 x i64>,<vscale x 2 x i64>)
+declare <vscale x 16 x i8> @llvm.aarch64.sve.bcax.nxv16i8(<vscale x 16 x i8>,<vscale x 16 x i8>,<vscale x 16 x i8>)
+declare <vscale x 8 x i16> @llvm.aarch64.sve.bcax.nxv8i16(<vscale x 8 x i16>,<vscale x 8 x i16>,<vscale x 8 x i16>)
+declare <vscale x 4 x i32> @llvm.aarch64.sve.bcax.nxv4i32(<vscale x 4 x i32>,<vscale x 4 x i32>,<vscale x 4 x i32>)
+declare <vscale x 2 x i64> @llvm.aarch64.sve.bcax.nxv2i64(<vscale x 2 x i64>,<vscale x 2 x i64>,<vscale x 2 x i64>)
+declare <vscale x 16 x i8> @llvm.aarch64.sve.bsl.nxv16i8(<vscale x 16 x i8>,<vscale x 16 x i8>,<vscale x 16 x i8>)
+declare <vscale x 8 x i16> @llvm.aarch64.sve.bsl.nxv8i16(<vscale x 8 x i16>,<vscale x 8 x i16>,<vscale x 8 x i16>)
+declare <vscale x 4 x i32> @llvm.aarch64.sve.bsl.nxv4i32(<vscale x 4 x i32>,<vscale x 4 x i32>,<vscale x 4 x i32>)
+declare <vscale x 2 x i64> @llvm.aarch64.sve.bsl.nxv2i64(<vscale x 2 x i64>,<vscale x 2 x i64>,<vscale x 2 x i64>)
+declare <vscale x 16 x i8> @llvm.aarch64.sve.bsl1n.nxv16i8(<vscale x 16 x i8>,<vscale x 16 x i8>,<vscale x 16 x i8>)
+declare <vscale x 8 x i16> @llvm.aarch64.sve.bsl1n.nxv8i16(<vscale x 8 x i16>,<vscale x 8 x i16>,<vscale x 8 x i16>)
+declare <vscale x 4 x i32> @llvm.aarch64.sve.bsl1n.nxv4i32(<vscale x 4 x i32>,<vscale x 4 x i32>,<vscale x 4 x i32>)
+declare <vscale x 2 x i64> @llvm.aarch64.sve.bsl1n.nxv2i64(<vscale x 2 x i64>,<vscale x 2 x i64>,<vscale x 2 x i64>)
+declare <vscale x 16 x i8> @llvm.aarch64.sve.bsl2n.nxv16i8(<vscale x 16 x i8>,<vscale x 16 x i8>,<vscale x 16 x i8>)
+declare <vscale x 8 x i16> @llvm.aarch64.sve.bsl2n.nxv8i16(<vscale x 8 x i16>,<vscale x 8 x i16>,<vscale x 8 x i16>)
+declare <vscale x 4 x i32> @llvm.aarch64.sve.bsl2n.nxv4i32(<vscale x 4 x i32>,<vscale x 4 x i32>,<vscale x 4 x i32>)
+declare <vscale x 2 x i64> @llvm.aarch64.sve.bsl2n.nxv2i64(<vscale x 2 x i64>,<vscale x 2 x i64>,<vscale x 2 x i64>)
+declare <vscale x 16 x i8> @llvm.aarch64.sve.nbsl.nxv16i8(<vscale x 16 x i8>,<vscale x 16 x i8>,<vscale x 16 x i8>)
+declare <vscale x 8 x i16> @llvm.aarch64.sve.nbsl.nxv8i16(<vscale x 8 x i16>,<vscale x 8 x i16>,<vscale x 8 x i16>)
+declare <vscale x 4 x i32> @llvm.aarch64.sve.nbsl.nxv4i32(<vscale x 4 x i32>,<vscale x 4 x i32>,<vscale x 4 x i32>)
+declare <vscale x 2 x i64> @llvm.aarch64.sve.nbsl.nxv2i64(<vscale x 2 x i64>,<vscale x 2 x i64>,<vscale x 2 x i64>)
-- 
2.34.1