[RISCV] Support k-ext clang intrinsics

[liaolucy]

The scalar crypto v1.0 builtins/intrinsics is still under discussion. Maybe we should wait for the final results？

[achieveartificialintelligence]

Rebase

[craig.topper]

clang/include/clang/Basic/BuiltinsRISCV.def

60

Capital Z

clang/lib/Basic/Targets/RISCV.cpp

189

This define doesn't seem very useful. It just tells you that you have some crypto instructions, but not which ones.

clang/lib/CodeGen/CGBuiltin.cpp

18709

Capital Z

[achieveartificialintelligence]

Address comments.

[achieveartificialintelligence]

Amend tests.

[achieveartificialintelligence]

Ping & Rebase.

[craig.topper]

clang/include/clang/Basic/BuiltinsRISCV.def

66

Uc -> IUc

The I will make the frontend only accept constant arguments.

73

Ui -> IUi

77

Uc -> IUc

100

Uc -> IUc

[achieveartificialintelligence]

Address @craig.topper's comments.

[achieveartificialintelligence]

Rebase

[achieveartificialintelligence]

Ping & Rebase

[craig.topper]

Any intrinsic that takes an immediate that needs to be in certain range, needs to have code added to SemaChecking.cpp to validate the range. Look for the other places we call SemaBuiltinConstantArgRange in Sema::CheckRISCVBuiltinFunctionCall

clang/include/clang/Basic/BuiltinsRISCV.def

62

Nothing is preventing __builtin_riscv_zip/unzip from being used on RV64 which will crash. We don't have a "32bit" feature flag so we can't fix it from here. You'll need to go into clang/lib/Sema/SemaChecking.cpp and implement something like where X86 emits err_32_bit_builtin_64_bit_tgt. Same is true for any other 32-bit only intrinsic.

[achieveartificialintelligence]

Thanks for @craig.topper's detailed advice!

[achieveartificialintelligence]

Fix typo

[craig.topper]

clang/include/clang/Basic/BuiltinsRISCV.def

61

Shouldn't brev8 be using "LiLi"?

62

Please add _32 to the end of zip and unzip.

clang/lib/Sema/SemaChecking.cpp

4092 ↗

(On Diff #411676)

What about aes64ks1i's rnum?

clang/test/CodeGen/RISCV/rvb-intrinsics/riscv32-zbkb.c

13 ↗

(On Diff #411676)

This should be long not int

clang/test/CodeGen/RISCV/rvb-intrinsics/riscv64-zbkb.c

1 ↗

(On Diff #411676)

Extra blank line at the top of the file

9 ↗

(On Diff #411676)

These should be i64

14 ↗

(On Diff #411676)

This should be long not int

[achieveartificialintelligence]

Address @craig.topper's comments. Thanks!

[craig.topper]

LGTM

clang/include/clang/Basic/BuiltinsRISCV.def

 TARGET_BUILTIN(__builtin_riscv_crc32_h, "LiLi", "nc", "experimental-zbr")
 TARGET_BUILTIN(__builtin_riscv_crc32_w, "LiLi", "nc", "experimental-zbr")
 TARGET_BUILTIN(__builtin_riscv_crc32c_b, "LiLi", "nc", "experimental-zbr")
 TARGET_BUILTIN(__builtin_riscv_crc32c_h, "LiLi", "nc", "experimental-zbr")
 TARGET_BUILTIN(__builtin_riscv_crc32c_w, "LiLi", "nc", "experimental-zbr")
 TARGET_BUILTIN(__builtin_riscv_crc32_d, "LiLi", "nc", "experimental-zbr")
 TARGET_BUILTIN(__builtin_riscv_crc32c_d, "LiLi", "nc", "experimental-zbr")
 
+// zbkb extension
+TARGET_BUILTIN(__builtin_riscv_brev8, "ZiZi", "nc", "experimental-zbkb")
+TARGET_BUILTIN(__builtin_riscv_zip, "ZiZi", "nc", "experimental-zbkb")
+TARGET_BUILTIN(__builtin_riscv_unzip, "ZiZi", "nc", "experimental-zbkb")
+
+// zbkc extension
+TARGET_BUILTIN(__builtin_riscv_k_clmul, "LiLiLi", "nc", "experimental-zbkc")
+TARGET_BUILTIN(__builtin_riscv_k_clmulh, "LiLiLi", "nc", "experimental-zbkc")
+
+// zbkx extension
+TARGET_BUILTIN(__builtin_riscv_xperm8, "LiLiLi", "nc", "experimental-zbkx")
+TARGET_BUILTIN(__builtin_riscv_xperm4, "LiLiLi", "nc", "experimental-zbkx")
+
+// zknd extension
+TARGET_BUILTIN(__builtin_riscv_aes32dsi, "LiLiLiUc", "nc", "experimental-zknd")
+TARGET_BUILTIN(__builtin_riscv_aes32dsmi, "ZiZiZiUc", "nc", "experimental-zknd")
+TARGET_BUILTIN(__builtin_riscv_aes64ds, "WiWiWi", "nc", "experimental-zknd")
+TARGET_BUILTIN(__builtin_riscv_aes64dsm, "WiWiWi", "nc", "experimental-zknd")
+TARGET_BUILTIN(__builtin_riscv_aes64im, "WiWi", "nc", "experimental-zknd")
+TARGET_BUILTIN(__builtin_riscv_aes64ks1i, "WiWiUi", "nc", "experimental-zknd")
+TARGET_BUILTIN(__builtin_riscv_aes64ks2, "WiWiWi", "nc", "experimental-zknd")
+
+// zkne extension
+TARGET_BUILTIN(__builtin_riscv_aes32esi, "ZiZiZiUc", "nc", "experimental-zkne")
+TARGET_BUILTIN(__builtin_riscv_aes32esmi, "ZiZiZiUc", "nc", "experimental-zkne")
+TARGET_BUILTIN(__builtin_riscv_aes64es, "WiWiWi", "nc", "experimental-zkne")
+TARGET_BUILTIN(__builtin_riscv_aes64esm, "WiWiWi", "nc", "experimental-zkne")
+
+// zknh extension
+TARGET_BUILTIN(__builtin_riscv_sha256sig0, "LiLi", "nc", "experimental-zknh")
+TARGET_BUILTIN(__builtin_riscv_sha256sig1, "LiLi", "nc", "experimental-zknh")
+TARGET_BUILTIN(__builtin_riscv_sha256sum0, "LiLi", "nc", "experimental-zknh")
+TARGET_BUILTIN(__builtin_riscv_sha256sum1, "LiLi", "nc", "experimental-zknh")
+TARGET_BUILTIN(__builtin_riscv_sha512sig0, "WiWi", "nc", "experimental-zknh")
+TARGET_BUILTIN(__builtin_riscv_sha512sig0h, "ZiZiZi", "nc", "experimental-zknh")
+TARGET_BUILTIN(__builtin_riscv_sha512sig0l, "ZiZiZi", "nc", "experimental-zknh")
+TARGET_BUILTIN(__builtin_riscv_sha512sig1, "WiWi", "nc", "experimental-zknh")
+TARGET_BUILTIN(__builtin_riscv_sha512sig1h, "ZiZiZi", "nc", "experimental-zknh")
+TARGET_BUILTIN(__builtin_riscv_sha512sig1l, "ZiZiZi", "nc", "experimental-zknh")
+TARGET_BUILTIN(__builtin_riscv_sha512sum0, "WiWi", "nc", "experimental-zknh")
+TARGET_BUILTIN(__builtin_riscv_sha512sum0r, "ZiZiZi", "nc", "experimental-zknh")
+TARGET_BUILTIN(__builtin_riscv_sha512sum1, "WiWi", "nc", "experimental-zknh")
+TARGET_BUILTIN(__builtin_riscv_sha512sum1r, "ZiZiZi", "nc", "experimental-zknh")
+
+// zkr extension
+TARGET_BUILTIN(__builtin_riscv_getnoise, "Li", "nc", "experimental-zkr")
+TARGET_BUILTIN(__builtin_riscv_pollentropy, "Li", "nc", "experimental-zkr")
+
+// zksh extension
+TARGET_BUILTIN(__builtin_riscv_sm3p0, "LiLi", "nc", "experimental-zksh")
+TARGET_BUILTIN(__builtin_riscv_sm3p1, "LiLi", "nc", "experimental-zksh")
+
+// zksed extension
+TARGET_BUILTIN(__builtin_riscv_sm4ed, "LiLiLiUc", "nc", "experimental-zksed")
+TARGET_BUILTIN(__builtin_riscv_sm4ks, "LiLiLiUc", "nc", "experimental-zksed")
+
 #undef BUILTIN
 #undef TARGET_BUILTIN

clang/lib/Basic/Targets/RISCV.cpp

if (FLen) {
     Builder.defineMacro("__riscv_fsqrt");
   }
 
   if (ISAInfo->hasExtension("c"))
     Builder.defineMacro("__riscv_compressed");
 
   if (ISAInfo->hasExtension("v"))
     Builder.defineMacro("__riscv_vector");
+
+  if (ISAInfo->hasExtension("zknd") || ISAInfo->hasExtension("zkne") ||
+      ISAInfo->hasExtension("zknh") || ISAInfo->hasExtension("zksh") ||
+      ISAInfo->hasExtension("zksed") || ISAInfo->hasExtension("zkr") ||
+      ISAInfo->hasExtension("zk"))
+    Builder.defineMacro("__riscv_crypto");
 }
 
 const Builtin::Info RISCVTargetInfo::BuiltinInfo[] = {
 #define BUILTIN(ID, TYPE, ATTRS)                                               \
   {#ID, TYPE, ATTRS, nullptr, ALL_LANGUAGES, nullptr},
 #define TARGET_BUILTIN(ID, TYPE, ATTRS, FEATURE)                               \
     {#ID, TYPE, ATTRS, nullptr, ALL_LANGUAGES, FEATURE},
 #include "clang/Basic/BuiltinsRISCVVector.def"

clang/lib/CodeGen/CGBuiltin.cpp

Value *CodeGenFunction::EmitRISCVBuiltinExpr(unsigned BuiltinID,
   case RISCV::BI__builtin_riscv_xperm_w:
   case RISCV::BI__builtin_riscv_crc32_b:
   case RISCV::BI__builtin_riscv_crc32_h:
   case RISCV::BI__builtin_riscv_crc32_w:
   case RISCV::BI__builtin_riscv_crc32_d:
   case RISCV::BI__builtin_riscv_crc32c_b:
   case RISCV::BI__builtin_riscv_crc32c_h:
   case RISCV::BI__builtin_riscv_crc32c_w:
-  case RISCV::BI__builtin_riscv_crc32c_d: {
+  case RISCV::BI__builtin_riscv_crc32c_d:
+  case RISCV::BI__builtin_riscv_brev8:
+  case RISCV::BI__builtin_riscv_zip:
+  case RISCV::BI__builtin_riscv_unzip:
+  case RISCV::BI__builtin_riscv_xperm8:
+  case RISCV::BI__builtin_riscv_xperm4:
+  case RISCV::BI__builtin_riscv_k_clmul:
+  case RISCV::BI__builtin_riscv_k_clmulh: {
     switch (BuiltinID) {
     default: llvm_unreachable("unexpected builtin ID");
     // Zbb
     case RISCV::BI__builtin_riscv_orc_b_32:
     case RISCV::BI__builtin_riscv_orc_b_64:
       ID = Intrinsic::riscv_orc_b;
       break;
 
case RISCV::BI__builtin_riscv_crc32c_h:
       ID = Intrinsic::riscv_crc32c_h;
       break;
     case RISCV::BI__builtin_riscv_crc32c_w:
       ID = Intrinsic::riscv_crc32c_w;
       break;
     case RISCV::BI__builtin_riscv_crc32c_d:
       ID = Intrinsic::riscv_crc32c_d;
       break;
+
+    // zbkb
+    case RISCV::BI__builtin_riscv_brev8:
+      ID = Intrinsic::riscv_brev8;
+      break;
+    case RISCV::BI__builtin_riscv_zip:
+      ID = Intrinsic::riscv_zip;
+      break;
+    case RISCV::BI__builtin_riscv_unzip:
+      ID = Intrinsic::riscv_unzip;
+      break;
+
+    // Zbkc
+    case RISCV::BI__builtin_riscv_k_clmul:
+      ID = Intrinsic::riscv_clmul;
+      break;
+    case RISCV::BI__builtin_riscv_k_clmulh:
+      ID = Intrinsic::riscv_clmulh;
+      break;
+
+    // zbkx
+    case RISCV::BI__builtin_riscv_xperm8:
+      ID = Intrinsic::riscv_xperm8;
+      break;
+    case RISCV::BI__builtin_riscv_xperm4:
+      ID = Intrinsic::riscv_xperm4;
+      break;
     }
 
     IntrinsicTypes = {ResultType};
     break;
   }
   // Vector builtins are handled from here.
 #include "clang/Basic/riscv_vector_builtin_cg.inc"
+
+  case RISCV::BI__builtin_riscv_aes32dsi:
+    ID = Intrinsic::riscv_aes32dsi;
+    break;
+  case RISCV::BI__builtin_riscv_aes32dsmi:
+    ID = Intrinsic::riscv_aes32dsmi;
+    break;
+  case RISCV::BI__builtin_riscv_aes32esi:
+    ID = Intrinsic::riscv_aes32esi;
+    break;
+  case RISCV::BI__builtin_riscv_aes32esmi:
+    ID = Intrinsic::riscv_aes32esmi;
+    break;
+  case RISCV::BI__builtin_riscv_aes64ds:
+    ID = Intrinsic::riscv_aes64ds;
+    break;
+  case RISCV::BI__builtin_riscv_aes64dsm:
+    ID = Intrinsic::riscv_aes64dsm;
+    break;
+  case RISCV::BI__builtin_riscv_aes64es:
+    ID = Intrinsic::riscv_aes64es;
+    break;
+  case RISCV::BI__builtin_riscv_aes64esm:
+    ID = Intrinsic::riscv_aes64esm;
+    break;
+  case RISCV::BI__builtin_riscv_aes64im:
+    ID = Intrinsic::riscv_aes64im;
+    break;
+  case RISCV::BI__builtin_riscv_aes64ks1i:
+    ID = Intrinsic::riscv_aes64ks1i;
+    break;
+  case RISCV::BI__builtin_riscv_aes64ks2:
+    ID = Intrinsic::riscv_aes64ks2;
+    break;
+  case RISCV::BI__builtin_riscv_getnoise:
+    ID = Intrinsic::riscv_getnoise;
+    IntrinsicTypes = {ResultType};
+    break;
+  case RISCV::BI__builtin_riscv_pollentropy:
+    ID = Intrinsic::riscv_pollentropy;
+    IntrinsicTypes = {ResultType};
+    break;
+  case RISCV::BI__builtin_riscv_sha256sig0:
+    ID = Intrinsic::riscv_sha256sig0;
+    IntrinsicTypes = {Ops[0]->getType(), ResultType};
+    break;
+  case RISCV::BI__builtin_riscv_sha256sig1:
+    ID = Intrinsic::riscv_sha256sig1;
+    IntrinsicTypes = {Ops[0]->getType(), ResultType};
+    break;
+  case RISCV::BI__builtin_riscv_sha256sum0:
+    ID = Intrinsic::riscv_sha256sum0;
+    IntrinsicTypes = {Ops[0]->getType(), ResultType};
+    break;
+  case RISCV::BI__builtin_riscv_sha256sum1:
+    ID = Intrinsic::riscv_sha256sum1;
+    IntrinsicTypes = {Ops[0]->getType(), ResultType};
+    break;
+  case RISCV::BI__builtin_riscv_sha512sig0:
+    ID = Intrinsic::riscv_sha512sig0;
+    break;
+  case RISCV::BI__builtin_riscv_sha512sig0h:
+    ID = Intrinsic::riscv_sha512sig0h;
+    break;
+  case RISCV::BI__builtin_riscv_sha512sig0l:
+    ID = Intrinsic::riscv_sha512sig0l;
+    break;
+  case RISCV::BI__builtin_riscv_sha512sig1:
+    ID = Intrinsic::riscv_sha512sig1;
+    break;
+  case RISCV::BI__builtin_riscv_sha512sig1h:
+    ID = Intrinsic::riscv_sha512sig1h;
+    break;
+  case RISCV::BI__builtin_riscv_sha512sig1l:
+    ID = Intrinsic::riscv_sha512sig1l;
+    break;
+  case RISCV::BI__builtin_riscv_sha512sum0:
+    ID = Intrinsic::riscv_sha512sum0;
+    break;
+  case RISCV::BI__builtin_riscv_sha512sum0r:
+    ID = Intrinsic::riscv_sha512sum0r;
+    break;
+  case RISCV::BI__builtin_riscv_sha512sum1:
+    ID = Intrinsic::riscv_sha512sum1;
+    break;
+  case RISCV::BI__builtin_riscv_sha512sum1r:
+    ID = Intrinsic::riscv_sha512sum1r;
+    break;
+  case RISCV::BI__builtin_riscv_sm3p0:
+    ID = Intrinsic::riscv_sm3p0;
+    IntrinsicTypes = {Ops[0]->getType(), ResultType};
+    break;
+  case RISCV::BI__builtin_riscv_sm3p1:
+    ID = Intrinsic::riscv_sm3p1;
+    IntrinsicTypes = {Ops[0]->getType(), ResultType};
+    break;
+  case RISCV::BI__builtin_riscv_sm4ed:
+    ID = Intrinsic::riscv_sm4ed;
+    IntrinsicTypes = {Ops[0]->getType(), Ops[1]->getType(), Ops[2]->getType(),
+                      ResultType};
+    break;
+  case RISCV::BI__builtin_riscv_sm4ks:
+    ID = Intrinsic::riscv_sm4ks;
+    IntrinsicTypes = {Ops[0]->getType(), Ops[1]->getType(), Ops[2]->getType(),
+                      ResultType};
+    break;
   }
 
   assert(ID != Intrinsic::not_intrinsic);
 
   llvm::Function *F = CGM.getIntrinsic(ID, IntrinsicTypes);
   return Builder.CreateCall(F, Ops, "");
 }

clang/test/CodeGen/RISCV/rvk-intrinsics/riscv32-zbkb.c

+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
+// RUN: %clang_cc1 -triple riscv32 -target-feature +experimental-zbkb -emit-llvm %s -o - \
+// RUN:     | FileCheck %s  -check-prefix=RV32ZBKB
+
+// RV32ZBKB-LABEL: @brev8(
+// RV32ZBKB-NEXT:  entry:
+// RV32ZBKB-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZBKB-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZBKB-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZBKB-NEXT:    [[TMP1:%.*]] = call i32 @llvm.riscv.brev8.i32(i32 [[TMP0]])
+// RV32ZBKB-NEXT:    ret i32 [[TMP1]]
+//
+int brev8(int rs1)
+{
+  return __builtin_riscv_brev8(rs1);
+}
+
+// RV32ZBKB-LABEL: @zip(
+// RV32ZBKB-NEXT:  entry:
+// RV32ZBKB-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZBKB-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZBKB-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZBKB-NEXT:    [[TMP1:%.*]] = call i32 @llvm.riscv.zip.i32(i32 [[TMP0]])
+// RV32ZBKB-NEXT:    ret i32 [[TMP1]]
+//
+int zip(int rs1)
+{
+  return __builtin_riscv_zip(rs1);
+}
+
+// RV32ZBKB-LABEL: @unzip(
+// RV32ZBKB-NEXT:  entry:
+// RV32ZBKB-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZBKB-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZBKB-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZBKB-NEXT:    [[TMP1:%.*]] = call i32 @llvm.riscv.unzip.i32(i32 [[TMP0]])
+// RV32ZBKB-NEXT:    ret i32 [[TMP1]]
+//
+int unzip(int rs1)
+{
+  return __builtin_riscv_unzip(rs1);
+}

clang/test/CodeGen/RISCV/rvk-intrinsics/riscv32-zbkc.c

+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
+// RUN: %clang_cc1 -triple riscv32 -target-feature +experimental-zbkc -emit-llvm %s -o - \
+// RUN:     | FileCheck %s  -check-prefix=RV32ZBKC
+
+// RV32ZBKC-LABEL: @clmul(
+// RV32ZBKC-NEXT:  entry:
+// RV32ZBKC-NEXT:    [[A_ADDR:%.*]] = alloca i32, align 4
+// RV32ZBKC-NEXT:    [[B_ADDR:%.*]] = alloca i32, align 4
+// RV32ZBKC-NEXT:    store i32 [[A:%.*]], i32* [[A_ADDR]], align 4
+// RV32ZBKC-NEXT:    store i32 [[B:%.*]], i32* [[B_ADDR]], align 4
+// RV32ZBKC-NEXT:    [[TMP0:%.*]] = load i32, i32* [[A_ADDR]], align 4
+// RV32ZBKC-NEXT:    [[TMP1:%.*]] = load i32, i32* [[B_ADDR]], align 4
+// RV32ZBKC-NEXT:    [[TMP2:%.*]] = call i32 @llvm.riscv.clmul.i32(i32 [[TMP0]], i32 [[TMP1]])
+// RV32ZBKC-NEXT:    ret i32 [[TMP2]]
+//
+long clmul(long a, long b) {
+  return __builtin_riscv_k_clmul(a, b);
+}
+
+// RV32ZBKC-LABEL: @clmulh(
+// RV32ZBKC-NEXT:  entry:
+// RV32ZBKC-NEXT:    [[A_ADDR:%.*]] = alloca i32, align 4
+// RV32ZBKC-NEXT:    [[B_ADDR:%.*]] = alloca i32, align 4
+// RV32ZBKC-NEXT:    store i32 [[A:%.*]], i32* [[A_ADDR]], align 4
+// RV32ZBKC-NEXT:    store i32 [[B:%.*]], i32* [[B_ADDR]], align 4
+// RV32ZBKC-NEXT:    [[TMP0:%.*]] = load i32, i32* [[A_ADDR]], align 4
+// RV32ZBKC-NEXT:    [[TMP1:%.*]] = load i32, i32* [[B_ADDR]], align 4
+// RV32ZBKC-NEXT:    [[TMP2:%.*]] = call i32 @llvm.riscv.clmulh.i32(i32 [[TMP0]], i32 [[TMP1]])
+// RV32ZBKC-NEXT:    ret i32 [[TMP2]]
+//
+long clmulh(long a, long b) {
+  return __builtin_riscv_k_clmulh(a, b);
+}

clang/test/CodeGen/RISCV/rvk-intrinsics/riscv32-zbkx.c

+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
+// RUN: %clang_cc1 -triple riscv32 -target-feature +experimental-zbkx -emit-llvm %s -o - \
+// RUN:     | FileCheck %s  -check-prefix=RV32ZBKX
+
+// RV32ZBKX-LABEL: @xperm8(
+// RV32ZBKX-NEXT:  entry:
+// RV32ZBKX-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZBKX-NEXT:    [[RS2_ADDR:%.*]] = alloca i32, align 4
+// RV32ZBKX-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZBKX-NEXT:    store i32 [[RS2:%.*]], i32* [[RS2_ADDR]], align 4
+// RV32ZBKX-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZBKX-NEXT:    [[TMP1:%.*]] = load i32, i32* [[RS2_ADDR]], align 4
+// RV32ZBKX-NEXT:    [[TMP2:%.*]] = call i32 @llvm.riscv.xperm8.i32(i32 [[TMP0]], i32 [[TMP1]])
+// RV32ZBKX-NEXT:    ret i32 [[TMP2]]
+//
+long xperm8(long rs1, long rs2)
+{
+  return __builtin_riscv_xperm8(rs1, rs2);
+}
+
+// RV32ZBKX-LABEL: @xperm4(
+// RV32ZBKX-NEXT:  entry:
+// RV32ZBKX-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZBKX-NEXT:    [[RS2_ADDR:%.*]] = alloca i32, align 4
+// RV32ZBKX-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZBKX-NEXT:    store i32 [[RS2:%.*]], i32* [[RS2_ADDR]], align 4
+// RV32ZBKX-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZBKX-NEXT:    [[TMP1:%.*]] = load i32, i32* [[RS2_ADDR]], align 4
+// RV32ZBKX-NEXT:    [[TMP2:%.*]] = call i32 @llvm.riscv.xperm4.i32(i32 [[TMP0]], i32 [[TMP1]])
+// RV32ZBKX-NEXT:    ret i32 [[TMP2]]
+//
+long xperm4(long rs1, long rs2)
+{
+  return __builtin_riscv_xperm4(rs1, rs2);
+}

clang/test/CodeGen/RISCV/rvk-intrinsics/riscv32-zknd.c

+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
+// RUN: %clang_cc1 -triple riscv32 -target-feature +experimental-zknd -emit-llvm %s -o - \
+// RUN:     | FileCheck %s  -check-prefix=RV32ZKND
+
+// RV32ZKND-LABEL: @aes32dsi(
+// RV32ZKND-NEXT:  entry:
+// RV32ZKND-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKND-NEXT:    [[RS2_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKND-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZKND-NEXT:    store i32 [[RS2:%.*]], i32* [[RS2_ADDR]], align 4
+// RV32ZKND-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZKND-NEXT:    [[TMP1:%.*]] = load i32, i32* [[RS2_ADDR]], align 4
+// RV32ZKND-NEXT:    [[TMP2:%.*]] = call i32 @llvm.riscv.aes32dsi(i32 [[TMP0]], i32 [[TMP1]], i8 3)
+// RV32ZKND-NEXT:    ret i32 [[TMP2]]
+//
+int aes32dsi(int rs1, int rs2) {
+  return __builtin_riscv_aes32dsi(rs1, rs2, 3);
+}
+
+// RV32ZKND-LABEL: @aes32dsmi(
+// RV32ZKND-NEXT:  entry:
+// RV32ZKND-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKND-NEXT:    [[RS2_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKND-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZKND-NEXT:    store i32 [[RS2:%.*]], i32* [[RS2_ADDR]], align 4
+// RV32ZKND-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZKND-NEXT:    [[TMP1:%.*]] = load i32, i32* [[RS2_ADDR]], align 4
+// RV32ZKND-NEXT:    [[TMP2:%.*]] = call i32 @llvm.riscv.aes32dsmi(i32 [[TMP0]], i32 [[TMP1]], i8 3)
+// RV32ZKND-NEXT:    ret i32 [[TMP2]]
+//
+int aes32dsmi(int rs1, int rs2) {
+  return __builtin_riscv_aes32dsmi(rs1, rs2, 3);
+}

clang/test/CodeGen/RISCV/rvk-intrinsics/riscv32-zkne.c

+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
+// RUN: %clang_cc1 -triple riscv32 -target-feature +experimental-zkne -emit-llvm %s -o - \
+// RUN:     | FileCheck %s  -check-prefix=RV32ZKNE
+
+// RV32ZKNE-LABEL: @aes32esi(
+// RV32ZKNE-NEXT:  entry:
+// RV32ZKNE-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKNE-NEXT:    [[RS2_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKNE-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZKNE-NEXT:    store i32 [[RS2:%.*]], i32* [[RS2_ADDR]], align 4
+// RV32ZKNE-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZKNE-NEXT:    [[TMP1:%.*]] = load i32, i32* [[RS2_ADDR]], align 4
+// RV32ZKNE-NEXT:    [[TMP2:%.*]] = call i32 @llvm.riscv.aes32esi(i32 [[TMP0]], i32 [[TMP1]], i8 3)
+// RV32ZKNE-NEXT:    ret i32 [[TMP2]]
+//
+int aes32esi(int rs1, int rs2) {
+  return __builtin_riscv_aes32esi(rs1, rs2, 3);
+}
+
+// RV32ZKNE-LABEL: @aes32esmi(
+// RV32ZKNE-NEXT:  entry:
+// RV32ZKNE-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKNE-NEXT:    [[RS2_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKNE-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZKNE-NEXT:    store i32 [[RS2:%.*]], i32* [[RS2_ADDR]], align 4
+// RV32ZKNE-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZKNE-NEXT:    [[TMP1:%.*]] = load i32, i32* [[RS2_ADDR]], align 4
+// RV32ZKNE-NEXT:    [[TMP2:%.*]] = call i32 @llvm.riscv.aes32esmi(i32 [[TMP0]], i32 [[TMP1]], i8 3)
+// RV32ZKNE-NEXT:    ret i32 [[TMP2]]
+//
+int aes32esmi(int rs1, int rs2) {
+  return __builtin_riscv_aes32esmi(rs1, rs2, 3);
+}

clang/test/CodeGen/RISCV/rvk-intrinsics/riscv32-zknh.c

+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
+// RUN: %clang_cc1 -triple riscv32 -target-feature +experimental-zknh -emit-llvm %s -o - \
+// RUN:     | FileCheck %s  -check-prefix=RV32ZKNH
+
+// RV32ZKNH-LABEL: @sha512sig0l(
+// RV32ZKNH-NEXT:  entry:
+// RV32ZKNH-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKNH-NEXT:    [[RS2_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKNH-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZKNH-NEXT:    store i32 [[RS2:%.*]], i32* [[RS2_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP1:%.*]] = load i32, i32* [[RS2_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP2:%.*]] = call i32 @llvm.riscv.sha512sig0l(i32 [[TMP0]], i32 [[TMP1]])
+// RV32ZKNH-NEXT:    ret i32 [[TMP2]]
+//
+int sha512sig0l(int rs1, int rs2) {
+  return __builtin_riscv_sha512sig0l(rs1, rs2);
+}
+
+
+// RV32ZKNH-LABEL: @sha512sig0h(
+// RV32ZKNH-NEXT:  entry:
+// RV32ZKNH-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKNH-NEXT:    [[RS2_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKNH-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZKNH-NEXT:    store i32 [[RS2:%.*]], i32* [[RS2_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP1:%.*]] = load i32, i32* [[RS2_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP2:%.*]] = call i32 @llvm.riscv.sha512sig0h(i32 [[TMP0]], i32 [[TMP1]])
+// RV32ZKNH-NEXT:    ret i32 [[TMP2]]
+//
+int sha512sig0h(int rs1, int rs2) {
+  return __builtin_riscv_sha512sig0h(rs1, rs2);
+}
+
+
+// RV32ZKNH-LABEL: @sha512sig1l(
+// RV32ZKNH-NEXT:  entry:
+// RV32ZKNH-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKNH-NEXT:    [[RS2_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKNH-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZKNH-NEXT:    store i32 [[RS2:%.*]], i32* [[RS2_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP1:%.*]] = load i32, i32* [[RS2_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP2:%.*]] = call i32 @llvm.riscv.sha512sig1l(i32 [[TMP0]], i32 [[TMP1]])
+// RV32ZKNH-NEXT:    ret i32 [[TMP2]]
+//
+int sha512sig1l(int rs1, int rs2) {
+  return __builtin_riscv_sha512sig1l(rs1, rs2);
+}
+
+// RV32ZKNH-LABEL: @sha512sig1h(
+// RV32ZKNH-NEXT:  entry:
+// RV32ZKNH-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKNH-NEXT:    [[RS2_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKNH-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZKNH-NEXT:    store i32 [[RS2:%.*]], i32* [[RS2_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP1:%.*]] = load i32, i32* [[RS2_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP2:%.*]] = call i32 @llvm.riscv.sha512sig1h(i32 [[TMP0]], i32 [[TMP1]])
+// RV32ZKNH-NEXT:    ret i32 [[TMP2]]
+//
+int sha512sig1h(int rs1, int rs2) {
+  return __builtin_riscv_sha512sig1h(rs1, rs2);
+}
+
+// RV32ZKNH-LABEL: @sha512sum0r(
+// RV32ZKNH-NEXT:  entry:
+// RV32ZKNH-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKNH-NEXT:    [[RS2_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKNH-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZKNH-NEXT:    store i32 [[RS2:%.*]], i32* [[RS2_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP1:%.*]] = load i32, i32* [[RS2_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP2:%.*]] = call i32 @llvm.riscv.sha512sum0r(i32 [[TMP0]], i32 [[TMP1]])
+// RV32ZKNH-NEXT:    ret i32 [[TMP2]]
+//
+int sha512sum0r(int rs1, int rs2) {
+  return __builtin_riscv_sha512sum0r(rs1, rs2);
+}
+
+
+// RV32ZKNH-LABEL: @sha512sum1r(
+// RV32ZKNH-NEXT:  entry:
+// RV32ZKNH-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKNH-NEXT:    [[RS2_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKNH-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZKNH-NEXT:    store i32 [[RS2:%.*]], i32* [[RS2_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP1:%.*]] = load i32, i32* [[RS2_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP2:%.*]] = call i32 @llvm.riscv.sha512sum1r(i32 [[TMP0]], i32 [[TMP1]])
+// RV32ZKNH-NEXT:    ret i32 [[TMP2]]
+//
+int sha512sum1r(int rs1, int rs2) {
+  return __builtin_riscv_sha512sum1r(rs1, rs2);
+}
+
+
+// RV32ZKNH-LABEL: @sha256sig0(
+// RV32ZKNH-NEXT:  entry:
+// RV32ZKNH-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKNH-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP1:%.*]] = call i32 @llvm.riscv.sha256sig0.i32.i32(i32 [[TMP0]])
+// RV32ZKNH-NEXT:    ret i32 [[TMP1]]
+//
+long sha256sig0(long rs1) {
+  return __builtin_riscv_sha256sig0(rs1);
+}
+
+
+// RV32ZKNH-LABEL: @sha256sig1(
+// RV32ZKNH-NEXT:  entry:
+// RV32ZKNH-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKNH-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP1:%.*]] = call i32 @llvm.riscv.sha256sig1.i32.i32(i32 [[TMP0]])
+// RV32ZKNH-NEXT:    ret i32 [[TMP1]]
+//
+long sha256sig1(long rs1) {
+  return __builtin_riscv_sha256sig1(rs1);
+}
+
+
+// RV32ZKNH-LABEL: @sha256sum0(
+// RV32ZKNH-NEXT:  entry:
+// RV32ZKNH-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKNH-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP1:%.*]] = call i32 @llvm.riscv.sha256sum0.i32.i32(i32 [[TMP0]])
+// RV32ZKNH-NEXT:    ret i32 [[TMP1]]
+//
+long sha256sum0(long rs1) {
+  return __builtin_riscv_sha256sum0(rs1);
+}
+
+
+// RV32ZKNH-LABEL: @sha256sum1(
+// RV32ZKNH-NEXT:  entry:
+// RV32ZKNH-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKNH-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZKNH-NEXT:    [[TMP1:%.*]] = call i32 @llvm.riscv.sha256sum1.i32.i32(i32 [[TMP0]])
+// RV32ZKNH-NEXT:    ret i32 [[TMP1]]
+//
+long sha256sum1(long rs1) {
+  return __builtin_riscv_sha256sum1(rs1);
+}

clang/test/CodeGen/RISCV/rvk-intrinsics/riscv32-zksed.c

+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
+// RUN: %clang_cc1 -triple riscv32 -target-feature +experimental-zksed -emit-llvm %s -o - \
+// RUN:     | FileCheck %s  -check-prefix=RV32ZKSED
+
+// RV32ZKSED-LABEL: @sm4ks(
+// RV32ZKSED-NEXT:  entry:
+// RV32ZKSED-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKSED-NEXT:    [[RS2_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKSED-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZKSED-NEXT:    store i32 [[RS2:%.*]], i32* [[RS2_ADDR]], align 4
+// RV32ZKSED-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZKSED-NEXT:    [[TMP1:%.*]] = load i32, i32* [[RS2_ADDR]], align 4
+// RV32ZKSED-NEXT:    [[TMP2:%.*]] = call i32 @llvm.riscv.sm4ks.i32.i32.i8.i32(i32 [[TMP0]], i32 [[TMP1]], i8 0)
+// RV32ZKSED-NEXT:    ret i32 [[TMP2]]
+//
+long sm4ks(long rs1, long rs2) {
+  return __builtin_riscv_sm4ks(rs1, rs2, 0);
+}
+
+
+// RV32ZKSED-LABEL: @sm4ed(
+// RV32ZKSED-NEXT:  entry:
+// RV32ZKSED-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKSED-NEXT:    [[RS2_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKSED-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZKSED-NEXT:    store i32 [[RS2:%.*]], i32* [[RS2_ADDR]], align 4
+// RV32ZKSED-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZKSED-NEXT:    [[TMP1:%.*]] = load i32, i32* [[RS2_ADDR]], align 4
+// RV32ZKSED-NEXT:    [[TMP2:%.*]] = call i32 @llvm.riscv.sm4ed.i32.i32.i8.i32(i32 [[TMP0]], i32 [[TMP1]], i8 0)
+// RV32ZKSED-NEXT:    ret i32 [[TMP2]]
+//
+long sm4ed(long rs1, long rs2) {
+  return __builtin_riscv_sm4ed(rs1, rs2, 0);
+}

clang/test/CodeGen/RISCV/rvk-intrinsics/riscv32-zksh.c

+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
+// RUN: %clang_cc1 -triple riscv32 -target-feature +experimental-zksh -emit-llvm %s -o - \
+// RUN:     | FileCheck %s  -check-prefix=RV32ZKSH
+
+// RV32ZKSH-LABEL: @sm3p0(
+// RV32ZKSH-NEXT:  entry:
+// RV32ZKSH-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKSH-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZKSH-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZKSH-NEXT:    [[TMP1:%.*]] = call i32 @llvm.riscv.sm3p0.i32.i32(i32 [[TMP0]])
+// RV32ZKSH-NEXT:    ret i32 [[TMP1]]
+//
+long sm3p0(long rs1)
+{
+  return __builtin_riscv_sm3p0(rs1);
+}
+
+// RV32ZKSH-LABEL: @sm3p1(
+// RV32ZKSH-NEXT:  entry:
+// RV32ZKSH-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV32ZKSH-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV32ZKSH-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV32ZKSH-NEXT:    [[TMP1:%.*]] = call i32 @llvm.riscv.sm3p1.i32.i32(i32 [[TMP0]])
+// RV32ZKSH-NEXT:    ret i32 [[TMP1]]
+//
+long sm3p1(long rs1) {
+  return __builtin_riscv_sm3p1(rs1);
+}

clang/test/CodeGen/RISCV/rvk-intrinsics/riscv64-zbkb.c

+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
+// RUN: %clang_cc1 -triple riscv64 -target-feature +experimental-zbkb -emit-llvm %s -o - \
+// RUN:     | FileCheck %s  -check-prefix=RV64ZBKB
+
+// RV64ZBKB-LABEL: @brev8(
+// RV64ZBKB-NEXT:  entry:
+// RV64ZBKB-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV64ZBKB-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV64ZBKB-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV64ZBKB-NEXT:    [[TMP1:%.*]] = call i32 @llvm.riscv.brev8.i32(i32 [[TMP0]])
+// RV64ZBKB-NEXT:    ret i32 [[TMP1]]
+//
+int brev8(int rs1)
+{
+  return __builtin_riscv_brev8(rs1);
+}

clang/test/CodeGen/RISCV/rvk-intrinsics/riscv64-zbkc.c

+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
+// RUN: %clang_cc1 -triple riscv64 -target-feature +experimental-zbkc -emit-llvm %s -o - \
+// RUN:     | FileCheck %s  -check-prefix=RV64ZBKC
+
+// RV64ZBKC-LABEL: @clmul(
+// RV64ZBKC-NEXT:  entry:
+// RV64ZBKC-NEXT:    [[A_ADDR:%.*]] = alloca i64, align 8
+// RV64ZBKC-NEXT:    [[B_ADDR:%.*]] = alloca i64, align 8
+// RV64ZBKC-NEXT:    store i64 [[A:%.*]], i64* [[A_ADDR]], align 8
+// RV64ZBKC-NEXT:    store i64 [[B:%.*]], i64* [[B_ADDR]], align 8
+// RV64ZBKC-NEXT:    [[TMP0:%.*]] = load i64, i64* [[A_ADDR]], align 8
+// RV64ZBKC-NEXT:    [[TMP1:%.*]] = load i64, i64* [[B_ADDR]], align 8
+// RV64ZBKC-NEXT:    [[TMP2:%.*]] = call i64 @llvm.riscv.clmul.i64(i64 [[TMP0]], i64 [[TMP1]])
+// RV64ZBKC-NEXT:    ret i64 [[TMP2]]
+//
+long clmul(long a, long b) {
+  return __builtin_riscv_k_clmul(a, b);
+}
+
+// RV64ZBKC-LABEL: @clmulh(
+// RV64ZBKC-NEXT:  entry:
+// RV64ZBKC-NEXT:    [[A_ADDR:%.*]] = alloca i64, align 8
+// RV64ZBKC-NEXT:    [[B_ADDR:%.*]] = alloca i64, align 8
+// RV64ZBKC-NEXT:    store i64 [[A:%.*]], i64* [[A_ADDR]], align 8
+// RV64ZBKC-NEXT:    store i64 [[B:%.*]], i64* [[B_ADDR]], align 8
+// RV64ZBKC-NEXT:    [[TMP0:%.*]] = load i64, i64* [[A_ADDR]], align 8
+// RV64ZBKC-NEXT:    [[TMP1:%.*]] = load i64, i64* [[B_ADDR]], align 8
+// RV64ZBKC-NEXT:    [[TMP2:%.*]] = call i64 @llvm.riscv.clmulh.i64(i64 [[TMP0]], i64 [[TMP1]])
+// RV64ZBKC-NEXT:    ret i64 [[TMP2]]
+//
+long clmulh(long a, long b) {
+  return __builtin_riscv_k_clmulh(a, b);
+}

clang/test/CodeGen/RISCV/rvk-intrinsics/riscv64-zbkx.c

+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
+// RUN: %clang_cc1 -triple riscv64 -target-feature +experimental-zbkx -emit-llvm %s -o - \
+// RUN:     | FileCheck %s  -check-prefix=RV64ZBKX
+
+// RV64ZBKX-LABEL: @xperm8(
+// RV64ZBKX-NEXT:  entry:
+// RV64ZBKX-NEXT:    [[RS1_ADDR:%.*]] = alloca i64, align 8
+// RV64ZBKX-NEXT:    [[RS2_ADDR:%.*]] = alloca i64, align 8
+// RV64ZBKX-NEXT:    store i64 [[RS1:%.*]], i64* [[RS1_ADDR]], align 8
+// RV64ZBKX-NEXT:    store i64 [[RS2:%.*]], i64* [[RS2_ADDR]], align 8
+// RV64ZBKX-NEXT:    [[TMP0:%.*]] = load i64, i64* [[RS1_ADDR]], align 8
+// RV64ZBKX-NEXT:    [[TMP1:%.*]] = load i64, i64* [[RS2_ADDR]], align 8
+// RV64ZBKX-NEXT:    [[TMP2:%.*]] = call i64 @llvm.riscv.xperm8.i64(i64 [[TMP0]], i64 [[TMP1]])
+// RV64ZBKX-NEXT:    ret i64 [[TMP2]]
+//
+long xperm8(long rs1, long rs2)
+{
+  return __builtin_riscv_xperm8(rs1, rs2);
+}
+
+// RV64ZBKX-LABEL: @xperm4(
+// RV64ZBKX-NEXT:  entry:
+// RV64ZBKX-NEXT:    [[RS1_ADDR:%.*]] = alloca i64, align 8
+// RV64ZBKX-NEXT:    [[RS2_ADDR:%.*]] = alloca i64, align 8
+// RV64ZBKX-NEXT:    store i64 [[RS1:%.*]], i64* [[RS1_ADDR]], align 8
+// RV64ZBKX-NEXT:    store i64 [[RS2:%.*]], i64* [[RS2_ADDR]], align 8
+// RV64ZBKX-NEXT:    [[TMP0:%.*]] = load i64, i64* [[RS1_ADDR]], align 8
+// RV64ZBKX-NEXT:    [[TMP1:%.*]] = load i64, i64* [[RS2_ADDR]], align 8
+// RV64ZBKX-NEXT:    [[TMP2:%.*]] = call i64 @llvm.riscv.xperm4.i64(i64 [[TMP0]], i64 [[TMP1]])
+// RV64ZBKX-NEXT:    ret i64 [[TMP2]]
+//
+long xperm4(long rs1, long rs2)
+{
+  return __builtin_riscv_xperm4(rs1, rs2);
+}

clang/test/CodeGen/RISCV/rvk-intrinsics/riscv64-zknd.c

+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
+// RUN: %clang_cc1 -triple riscv64 -target-feature +experimental-zknd -emit-llvm %s -o - \
+// RUN:     | FileCheck %s  -check-prefix=RV64ZKND
+
+
+// RV64ZKND-LABEL: @es64dsm(
+// RV64ZKND-NEXT:  entry:
+// RV64ZKND-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV64ZKND-NEXT:    [[RS2_ADDR:%.*]] = alloca i32, align 4
+// RV64ZKND-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV64ZKND-NEXT:    store i32 [[RS2:%.*]], i32* [[RS2_ADDR]], align 4
+// RV64ZKND-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV64ZKND-NEXT:    [[CONV:%.*]] = sext i32 [[TMP0]] to i64
+// RV64ZKND-NEXT:    [[TMP1:%.*]] = load i32, i32* [[RS2_ADDR]], align 4
+// RV64ZKND-NEXT:    [[CONV1:%.*]] = sext i32 [[TMP1]] to i64
+// RV64ZKND-NEXT:    [[TMP2:%.*]] = call i64 @llvm.riscv.aes64dsm(i64 [[CONV]], i64 [[CONV1]])
+// RV64ZKND-NEXT:    [[CONV2:%.*]] = trunc i64 [[TMP2]] to i32
+// RV64ZKND-NEXT:    ret i32 [[CONV2]]
+//
+int es64dsm(int rs1, int rs2) {
+  return __builtin_riscv_aes64dsm(rs1, rs2);
+}
+
+
+// RV64ZKND-LABEL: @es64ds(
+// RV64ZKND-NEXT:  entry:
+// RV64ZKND-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV64ZKND-NEXT:    [[RS2_ADDR:%.*]] = alloca i32, align 4
+// RV64ZKND-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV64ZKND-NEXT:    store i32 [[RS2:%.*]], i32* [[RS2_ADDR]], align 4
+// RV64ZKND-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV64ZKND-NEXT:    [[CONV:%.*]] = sext i32 [[TMP0]] to i64
+// RV64ZKND-NEXT:    [[TMP1:%.*]] = load i32, i32* [[RS2_ADDR]], align 4
+// RV64ZKND-NEXT:    [[CONV1:%.*]] = sext i32 [[TMP1]] to i64
+// RV64ZKND-NEXT:    [[TMP2:%.*]] = call i64 @llvm.riscv.aes64ds(i64 [[CONV]], i64 [[CONV1]])
+// RV64ZKND-NEXT:    [[CONV2:%.*]] = trunc i64 [[TMP2]] to i32
+// RV64ZKND-NEXT:    ret i32 [[CONV2]]
+//
+int es64ds(int rs1, int rs2) {
+  return __builtin_riscv_aes64ds(rs1, rs2);
+}
+
+
+// RV64ZKND-LABEL: @es64im(
+// RV64ZKND-NEXT:  entry:
+// RV64ZKND-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV64ZKND-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV64ZKND-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV64ZKND-NEXT:    [[CONV:%.*]] = sext i32 [[TMP0]] to i64
+// RV64ZKND-NEXT:    [[TMP1:%.*]] = call i64 @llvm.riscv.aes64im(i64 [[CONV]])
+// RV64ZKND-NEXT:    [[CONV1:%.*]] = trunc i64 [[TMP1]] to i32
+// RV64ZKND-NEXT:    ret i32 [[CONV1]]
+//
+int es64im(int rs1) {
+  return __builtin_riscv_aes64im(rs1);
+}
+
+// RV64ZKND-LABEL: @aes64ks1i(
+// RV64ZKND-NEXT:  entry:
+// RV64ZKND-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV64ZKND-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV64ZKND-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV64ZKND-NEXT:    [[CONV:%.*]] = sext i32 [[TMP0]] to i64
+// RV64ZKND-NEXT:    [[TMP1:%.*]] = call i64 @llvm.riscv.aes64ks1i(i64 [[CONV]], i32 0)
+// RV64ZKND-NEXT:    [[CONV1:%.*]] = trunc i64 [[TMP1]] to i32
+// RV64ZKND-NEXT:    ret i32 [[CONV1]]
+//
+int aes64ks1i(int rs1) {
+  return __builtin_riscv_aes64ks1i(rs1, 0);
+}
+
+// RV64ZKND-LABEL: @aes64ks2(
+// RV64ZKND-NEXT:  entry:
+// RV64ZKND-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV64ZKND-NEXT:    [[RS2_ADDR:%.*]] = alloca i32, align 4
+// RV64ZKND-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV64ZKND-NEXT:    store i32 [[RS2:%.*]], i32* [[RS2_ADDR]], align 4
+// RV64ZKND-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV64ZKND-NEXT:    [[CONV:%.*]] = sext i32 [[TMP0]] to i64
+// RV64ZKND-NEXT:    [[TMP1:%.*]] = load i32, i32* [[RS2_ADDR]], align 4
+// RV64ZKND-NEXT:    [[CONV1:%.*]] = sext i32 [[TMP1]] to i64
+// RV64ZKND-NEXT:    [[TMP2:%.*]] = call i64 @llvm.riscv.aes64ks2(i64 [[CONV]], i64 [[CONV1]])
+// RV64ZKND-NEXT:    [[CONV2:%.*]] = trunc i64 [[TMP2]] to i32
+// RV64ZKND-NEXT:    ret i32 [[CONV2]]
+//
+int aes64ks2(int rs1, int rs2) {
+  return __builtin_riscv_aes64ks2(rs1, rs2);
+}

clang/test/CodeGen/RISCV/rvk-intrinsics/riscv64-zkne.c

+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
+// RUN: %clang_cc1 -triple riscv64 -target-feature +experimental-zkne -emit-llvm %s -o - \
+// RUN:     | FileCheck %s  -check-prefix=RV64ZKNE
+
+
+// RV64ZKNE-LABEL: @aes64es(
+// RV64ZKNE-NEXT:  entry:
+// RV64ZKNE-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV64ZKNE-NEXT:    [[RS2_ADDR:%.*]] = alloca i32, align 4
+// RV64ZKNE-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV64ZKNE-NEXT:    store i32 [[RS2:%.*]], i32* [[RS2_ADDR]], align 4
+// RV64ZKNE-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV64ZKNE-NEXT:    [[CONV:%.*]] = sext i32 [[TMP0]] to i64
+// RV64ZKNE-NEXT:    [[TMP1:%.*]] = load i32, i32* [[RS2_ADDR]], align 4
+// RV64ZKNE-NEXT:    [[CONV1:%.*]] = sext i32 [[TMP1]] to i64
+// RV64ZKNE-NEXT:    [[TMP2:%.*]] = call i64 @llvm.riscv.aes64es(i64 [[CONV]], i64 [[CONV1]])
+// RV64ZKNE-NEXT:    [[CONV2:%.*]] = trunc i64 [[TMP2]] to i32
+// RV64ZKNE-NEXT:    ret i32 [[CONV2]]
+//
+int aes64es(int rs1, int rs2) {
+  return __builtin_riscv_aes64es(rs1, rs2);
+}
+
+
+// RV64ZKNE-LABEL: @aes64esm(
+// RV64ZKNE-NEXT:  entry:
+// RV64ZKNE-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV64ZKNE-NEXT:    [[RS2_ADDR:%.*]] = alloca i32, align 4
+// RV64ZKNE-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV64ZKNE-NEXT:    store i32 [[RS2:%.*]], i32* [[RS2_ADDR]], align 4
+// RV64ZKNE-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV64ZKNE-NEXT:    [[CONV:%.*]] = sext i32 [[TMP0]] to i64
+// RV64ZKNE-NEXT:    [[TMP1:%.*]] = load i32, i32* [[RS2_ADDR]], align 4
+// RV64ZKNE-NEXT:    [[CONV1:%.*]] = sext i32 [[TMP1]] to i64
+// RV64ZKNE-NEXT:    [[TMP2:%.*]] = call i64 @llvm.riscv.aes64esm(i64 [[CONV]], i64 [[CONV1]])
+// RV64ZKNE-NEXT:    [[CONV2:%.*]] = trunc i64 [[TMP2]] to i32
+// RV64ZKNE-NEXT:    ret i32 [[CONV2]]
+//
+int aes64esm(int rs1, int rs2) {
+  return __builtin_riscv_aes64esm(rs1, rs2);
+}

clang/test/CodeGen/RISCV/rvk-intrinsics/riscv64-zknh.c

+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
+// RUN: %clang_cc1 -triple riscv64 -target-feature +experimental-zknh -emit-llvm %s -o - \
+// RUN:     | FileCheck %s  -check-prefix=RV64ZKNH
+
+
+// RV64ZKNH-LABEL: @sha512sig0(
+// RV64ZKNH-NEXT:  entry:
+// RV64ZKNH-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV64ZKNH-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV64ZKNH-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV64ZKNH-NEXT:    [[CONV:%.*]] = sext i32 [[TMP0]] to i64
+// RV64ZKNH-NEXT:    [[TMP1:%.*]] = call i64 @llvm.riscv.sha512sig0(i64 [[CONV]])
+// RV64ZKNH-NEXT:    [[CONV1:%.*]] = trunc i64 [[TMP1]] to i32
+// RV64ZKNH-NEXT:    ret i32 [[CONV1]]
+//
+int sha512sig0(int rs1) {
+  return __builtin_riscv_sha512sig0(rs1);
+}
+
+
+// RV64ZKNH-LABEL: @sha512sig1(
+// RV64ZKNH-NEXT:  entry:
+// RV64ZKNH-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV64ZKNH-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV64ZKNH-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV64ZKNH-NEXT:    [[CONV:%.*]] = sext i32 [[TMP0]] to i64
+// RV64ZKNH-NEXT:    [[TMP1:%.*]] = call i64 @llvm.riscv.sha512sig1(i64 [[CONV]])
+// RV64ZKNH-NEXT:    [[CONV1:%.*]] = trunc i64 [[TMP1]] to i32
+// RV64ZKNH-NEXT:    ret i32 [[CONV1]]
+//
+int sha512sig1(int rs1) {
+  return __builtin_riscv_sha512sig1(rs1);
+}
+
+
+// RV64ZKNH-LABEL: @sha512sum0(
+// RV64ZKNH-NEXT:  entry:
+// RV64ZKNH-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV64ZKNH-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV64ZKNH-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV64ZKNH-NEXT:    [[CONV:%.*]] = sext i32 [[TMP0]] to i64
+// RV64ZKNH-NEXT:    [[TMP1:%.*]] = call i64 @llvm.riscv.sha512sum0(i64 [[CONV]])
+// RV64ZKNH-NEXT:    [[CONV1:%.*]] = trunc i64 [[TMP1]] to i32
+// RV64ZKNH-NEXT:    ret i32 [[CONV1]]
+//
+int sha512sum0(int rs1) {
+  return __builtin_riscv_sha512sum0(rs1);
+}
+
+
+// RV64ZKNH-LABEL: @sha512sum1(
+// RV64ZKNH-NEXT:  entry:
+// RV64ZKNH-NEXT:    [[RS1_ADDR:%.*]] = alloca i32, align 4
+// RV64ZKNH-NEXT:    store i32 [[RS1:%.*]], i32* [[RS1_ADDR]], align 4
+// RV64ZKNH-NEXT:    [[TMP0:%.*]] = load i32, i32* [[RS1_ADDR]], align 4
+// RV64ZKNH-NEXT:    [[CONV:%.*]] = sext i32 [[TMP0]] to i64
+// RV64ZKNH-NEXT:    [[TMP1:%.*]] = call i64 @llvm.riscv.sha512sum1(i64 [[CONV]])
+// RV64ZKNH-NEXT:    [[CONV1:%.*]] = trunc i64 [[TMP1]] to i32
+// RV64ZKNH-NEXT:    ret i32 [[CONV1]]
+//
+int sha512sum1(int rs1) {
+  return __builtin_riscv_sha512sum1(rs1);
+}
+
+
+// RV64ZKNH-LABEL: @sha256sig0(
+// RV64ZKNH-NEXT:  entry:
+// RV64ZKNH-NEXT:    [[RS1_ADDR:%.*]] = alloca i64, align 8
+// RV64ZKNH-NEXT:    store i64 [[RS1:%.*]], i64* [[RS1_ADDR]], align 8
+// RV64ZKNH-NEXT:    [[TMP0:%.*]] = load i64, i64* [[RS1_ADDR]], align 8
+// RV64ZKNH-NEXT:    [[TMP1:%.*]] = call i64 @llvm.riscv.sha256sig0.i64.i64(i64 [[TMP0]])
+// RV64ZKNH-NEXT:    ret i64 [[TMP1]]
+//
+long sha256sig0(long rs1) {
+  return __builtin_riscv_sha256sig0(rs1);
+}
+
+// RV64ZKNH-LABEL: @sha256sig1(
+// RV64ZKNH-NEXT:  entry:
+// RV64ZKNH-NEXT:    [[RS1_ADDR:%.*]] = alloca i64, align 8
+// RV64ZKNH-NEXT:    store i64 [[RS1:%.*]], i64* [[RS1_ADDR]], align 8
+// RV64ZKNH-NEXT:    [[TMP0:%.*]] = load i64, i64* [[RS1_ADDR]], align 8
+// RV64ZKNH-NEXT:    [[TMP1:%.*]] = call i64 @llvm.riscv.sha256sig1.i64.i64(i64 [[TMP0]])
+// RV64ZKNH-NEXT:    ret i64 [[TMP1]]
+//
+long sha256sig1(long rs1) {
+  return __builtin_riscv_sha256sig1(rs1);
+}
+
+
+// RV64ZKNH-LABEL: @sha256sum0(
+// RV64ZKNH-NEXT:  entry:
+// RV64ZKNH-NEXT:    [[RS1_ADDR:%.*]] = alloca i64, align 8
+// RV64ZKNH-NEXT:    store i64 [[RS1:%.*]], i64* [[RS1_ADDR]], align 8
+// RV64ZKNH-NEXT:    [[TMP0:%.*]] = load i64, i64* [[RS1_ADDR]], align 8
+// RV64ZKNH-NEXT:    [[TMP1:%.*]] = call i64 @llvm.riscv.sha256sum0.i64.i64(i64 [[TMP0]])
+// RV64ZKNH-NEXT:    ret i64 [[TMP1]]
+//
+long sha256sum0(long rs1) {
+  return __builtin_riscv_sha256sum0(rs1);
+}
+
+// RV64ZKNH-LABEL: @sha256sum1(
+// RV64ZKNH-NEXT:  entry:
+// RV64ZKNH-NEXT:    [[RS1_ADDR:%.*]] = alloca i64, align 8
+// RV64ZKNH-NEXT:    store i64 [[RS1:%.*]], i64* [[RS1_ADDR]], align 8
+// RV64ZKNH-NEXT:    [[TMP0:%.*]] = load i64, i64* [[RS1_ADDR]], align 8
+// RV64ZKNH-NEXT:    [[TMP1:%.*]] = call i64 @llvm.riscv.sha256sum1.i64.i64(i64 [[TMP0]])
+// RV64ZKNH-NEXT:    ret i64 [[TMP1]]
+//
+long sha256sum1(long rs1) {
+  return __builtin_riscv_sha256sum1(rs1);
+}

clang/test/CodeGen/RISCV/rvk-intrinsics/riscv64-zksed.c

+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
+// RUN: %clang_cc1 -triple riscv64 -target-feature +experimental-zksed -emit-llvm %s -o - \
+// RUN:     | FileCheck %s  -check-prefix=RV64ZKSED
+
+// RV64ZKSED-LABEL: @sm4ks(
+// RV64ZKSED-NEXT:  entry:
+// RV64ZKSED-NEXT:    [[RS1_ADDR:%.*]] = alloca i64, align 8
+// RV64ZKSED-NEXT:    [[RS2_ADDR:%.*]] = alloca i64, align 8
+// RV64ZKSED-NEXT:    store i64 [[RS1:%.*]], i64* [[RS1_ADDR]], align 8
+// RV64ZKSED-NEXT:    store i64 [[RS2:%.*]], i64* [[RS2_ADDR]], align 8
+// RV64ZKSED-NEXT:    [[TMP0:%.*]] = load i64, i64* [[RS1_ADDR]], align 8
+// RV64ZKSED-NEXT:    [[TMP1:%.*]] = load i64, i64* [[RS2_ADDR]], align 8
+// RV64ZKSED-NEXT:    [[TMP2:%.*]] = call i64 @llvm.riscv.sm4ks.i64.i64.i8.i64(i64 [[TMP0]], i64 [[TMP1]], i8 0)
+// RV64ZKSED-NEXT:    ret i64 [[TMP2]]
+//
+long sm4ks(long rs1, long rs2) {
+  return __builtin_riscv_sm4ks(rs1, rs2, 0);
+}
+
+
+// RV64ZKSED-LABEL: @sm4ed(
+// RV64ZKSED-NEXT:  entry:
+// RV64ZKSED-NEXT:    [[RS1_ADDR:%.*]] = alloca i64, align 8
+// RV64ZKSED-NEXT:    [[RS2_ADDR:%.*]] = alloca i64, align 8
+// RV64ZKSED-NEXT:    store i64 [[RS1:%.*]], i64* [[RS1_ADDR]], align 8
+// RV64ZKSED-NEXT:    store i64 [[RS2:%.*]], i64* [[RS2_ADDR]], align 8
+// RV64ZKSED-NEXT:    [[TMP0:%.*]] = load i64, i64* [[RS1_ADDR]], align 8
+// RV64ZKSED-NEXT:    [[TMP1:%.*]] = load i64, i64* [[RS2_ADDR]], align 8
+// RV64ZKSED-NEXT:    [[TMP2:%.*]] = call i64 @llvm.riscv.sm4ed.i64.i64.i8.i64(i64 [[TMP0]], i64 [[TMP1]], i8 0)
+// RV64ZKSED-NEXT:    ret i64 [[TMP2]]
+//
+long sm4ed(long rs1, long rs2) {
+  return __builtin_riscv_sm4ed(rs1, rs2, 0);
+}

clang/test/CodeGen/RISCV/rvk-intrinsics/riscv64-zksh.c

+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
+// RUN: %clang_cc1 -triple riscv64 -target-feature +experimental-zksh -emit-llvm %s -o - \
+// RUN:     | FileCheck %s  -check-prefix=RV64ZKSH
+
+// RV64ZKSH-LABEL: @sm3p0(
+// RV64ZKSH-NEXT:  entry:
+// RV64ZKSH-NEXT:    [[RS1_ADDR:%.*]] = alloca i64, align 8
+// RV64ZKSH-NEXT:    store i64 [[RS1:%.*]], i64* [[RS1_ADDR]], align 8
+// RV64ZKSH-NEXT:    [[TMP0:%.*]] = load i64, i64* [[RS1_ADDR]], align 8
+// RV64ZKSH-NEXT:    [[TMP1:%.*]] = call i64 @llvm.riscv.sm3p0.i64.i64(i64 [[TMP0]])
+// RV64ZKSH-NEXT:    ret i64 [[TMP1]]
+//
+long sm3p0(long rs1) {
+  return __builtin_riscv_sm3p0(rs1);
+}
+
+
+// RV64ZKSH-LABEL: @sm3p1(
+// RV64ZKSH-NEXT:  entry:
+// RV64ZKSH-NEXT:    [[RS1_ADDR:%.*]] = alloca i64, align 8
+// RV64ZKSH-NEXT:    store i64 [[RS1:%.*]], i64* [[RS1_ADDR]], align 8
+// RV64ZKSH-NEXT:    [[TMP0:%.*]] = load i64, i64* [[RS1_ADDR]], align 8
+// RV64ZKSH-NEXT:    [[TMP1:%.*]] = call i64 @llvm.riscv.sm3p1.i64.i64(i64 [[TMP0]])
+// RV64ZKSH-NEXT:    ret i64 [[TMP1]]
+//
+long sm3p1(long rs1) {
+  return __builtin_riscv_sm3p1(rs1);
+}

clang/test/Driver/riscv-arch.c

 // RUN: %clang -target riscv32-unknown-elf -march=rv32izbb1p0zbp0p93 -menable-experimental-extensions -### %s \
 // RUN: -fsyntax-only 2>&1 | FileCheck -check-prefix=RV32-EXPERIMENTAL-ZBB-ZBP-UNDERSCORE %s
 // RV32-EXPERIMENTAL-ZBB-ZBP-UNDERSCORE: error: invalid arch name 'rv32izbb1p0zbp0p93', multi-character extensions must be separated by underscores
 
 // RUN: %clang -target riscv32-unknown-elf -march=rv32izba1p0 -menable-experimental-extensions -### %s \
 // RUN: -fsyntax-only 2>&1 | FileCheck -check-prefix=RV32-EXPERIMENTAL-ZBA %s
 // RV32-EXPERIMENTAL-ZBA: "-target-feature" "+experimental-zba"
 
+// RUN: %clang -target riscv32-unknown-elf -march=rv32izbkb -### %s \
+// RUN: -fsyntax-only 2>&1 | FileCheck -check-prefix=RV32-EXPERIMENTAL-ZBKB-NOFLAG %s
+// RV32-EXPERIMENTAL-ZBKB-NOFLAG: error: invalid arch name 'rv32izbkb'
+// RV32-EXPERIMENTAL-ZBKB-NOFLAG: requires '-menable-experimental-extensions'
+
+// RUN: %clang -target riscv32-unknown-elf -march=rv32izbkb1p0 -menable-experimental-extensions -### %s \
+// RUN: -fsyntax-only 2>&1 | FileCheck -check-prefix=RV32-EXPERIMENTAL-ZBKB %s
+// RV32-EXPERIMENTAL-ZBKB: "-target-feature" "+experimental-zbkb"
+
 // RUN: %clang -target riscv32-unknown-elf -march=rv32iv -### %s -c 2>&1 | \
 // RUN:   FileCheck -check-prefix=RV32-EXPERIMENTAL-V-NOFLAG %s
 // RV32-EXPERIMENTAL-V-NOFLAG: error: invalid arch name 'rv32iv'
 // RV32-EXPERIMENTAL-V-NOFLAG: requires '-menable-experimental-extensions'
 
 // RUN: %clang -target riscv32-unknown-elf -march=rv32iv -menable-experimental-extensions -### %s -c 2>&1 | \
 // RUN:   FileCheck -check-prefix=RV32-EXPERIMENTAL-V-NOVERS %s
 // RV32-EXPERIMENTAL-V-NOVERS: error: invalid arch name 'rv32iv'

clang/test/Preprocessor/riscv-target-features.c

 // CHECK-NOT: __riscv_zbc
 // CHECK-NOT: __riscv_zbe
 // CHECK-NOT: __riscv_zbf
 // CHECK-NOT: __riscv_zbm
 // CHECK-NOT: __riscv_zbp
 // CHECK-NOT: __riscv_zbr
 // CHECK-NOT: __riscv_zbs
 // CHECK-NOT: __riscv_zbt
+// CHECK-NOT: __riscv_zbkb
+// CHECK-NOT: __riscv_zbkc
+// CHECK-NOT: __riscv_zbkx
+// CHECK-NOT: __riscv_zknd
+// CHECK-NOT: __riscv_zkne
+// CHECK-NOT: __riscv_zknh
+// CHECK-NOT: __riscv_zksed
+// CHECK-NOT: __riscv_zksh
+// CHECK-NOT: __riscv_zkr
+// CHECK-NOT: __riscv_zkn
+// CHECK-NOT: __riscv_zks
+// CHECK-NOT: __riscv_zkt
+// CHECK-NOT: __riscv_zk
 // CHECK-NOT: __riscv_zfh
 // CHECK-NOT: __riscv_v
 // CHECK-NOT: __riscv_vector
 // CHECK-NOT: __riscv_zvamo
 // CHECK-NOT: __riscv_zvlsseg
 
 // RUN: %clang -target riscv32-unknown-linux-gnu -march=rv32im -x c -E -dM %s \
 // RUN: -o - | FileCheck --check-prefix=CHECK-M-EXT %s

 // RUN: -o - | FileCheck --check-prefix=CHECK-ZBT-EXT %s
 // RUN: %clang -target riscv64-unknown-linux-gnu -menable-experimental-extensions \
 // RUN: -march=rv64izbt0p93 -x c -E -dM %s \
 // RUN: -o - | FileCheck --check-prefix=CHECK-ZBT-EXT %s
 // CHECK-ZBT-NOT: __riscv_b
 // CHECK-ZBT-EXT: __riscv_zbt 93000
 
 // RUN: %clang -target riscv32-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv32izbkb1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZBKB-EXT %s
+// RUN: %clang -target riscv64-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv64izbkb1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZBKB-EXT %s
+// CHECK-ZBKB-NOT: __riscv_k
+// CHECK-ZBKB-EXT: __riscv_zbkb 100000
+
+// RUN: %clang -target riscv32-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv32izbkc1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZBKC-EXT %s
+// RUN: %clang -target riscv64-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv64izbkc1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZBKC-EXT %s
+// CHECK-ZBKC-NOT: __riscv_k
+// CHECK-ZBKC-EXT: __riscv_zbkc 100000
+
+// RUN: %clang -target riscv32-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv32izbkx1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZBKX-EXT %s
+// RUN: %clang -target riscv64-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv64izbkx1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZBKX-EXT %s
+// CHECK-ZBKX-NOT: __riscv_k
+// CHECK-ZBKX-EXT: __riscv_zbkx 100000
+
+// RUN: %clang -target riscv32-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv32izknd1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZKND-EXT %s
+// RUN: %clang -target riscv64-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv64izknd1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZKND-EXT %s
+// CHECK-ZKND-NOT: __riscv_k
+// CHECK-ZKND-EXT: __riscv_zknd 100000
+
+// RUN: %clang -target riscv32-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv32izkne1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZKNE-EXT %s
+// RUN: %clang -target riscv64-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv64izkne1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZKNE-EXT %s
+// CHECK-ZKNE-NOT: __riscv_k
+// CHECK-ZKNE-EXT: __riscv_zkne 100000
+
+// RUN: %clang -target riscv32-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv32izknh1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZKNH-EXT %s
+// RUN: %clang -target riscv64-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv64izknh1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZKNH-EXT %s
+// CHECK-ZKNH-NOT: __riscv_k
+// CHECK-ZKNH-EXT: __riscv_zknh 100000
+
+// RUN: %clang -target riscv32-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv32izksed1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZKSED-EXT %s
+// RUN: %clang -target riscv64-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv64izksed1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZKSED-EXT %s
+// CHECK-ZKSED-NOT: __riscv_k
+// CHECK-ZKSED-EXT: __riscv_zksed 100000
+
+// RUN: %clang -target riscv32-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv32izksh1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZKSH-EXT %s
+// RUN: %clang -target riscv64-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv64izksh1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZKSH-EXT %s
+// CHECK-ZKSH-NOT: __riscv_k
+// CHECK-ZKSH-EXT: __riscv_zksh 100000
+
+// RUN: %clang -target riscv32-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv32izkr1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZKR-EXT %s
+// RUN: %clang -target riscv64-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv64izkr1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZKR-EXT %s
+// CHECK-ZKR-NOT: __riscv_k
+// CHECK-ZKR-EXT: __riscv_zkr 100000
+
+// RUN: %clang -target riscv32-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv32izkn1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZKN-EXT %s
+// RUN: %clang -target riscv64-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv64izkn1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZKN-EXT %s
+// CHECK-ZKN-NOT: __riscv_k
+// CHECK-ZKN-EXT: __riscv_zkn 100000
+
+// RUN: %clang -target riscv32-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv32izks1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZKS-EXT %s
+// RUN: %clang -target riscv64-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv64izks1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZKS-EXT %s
+// CHECK-ZKS-NOT: __riscv_k
+// CHECK-ZKS-EXT: __riscv_zks 100000
+
+// RUN: %clang -target riscv32-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv32izkt1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZKT-EXT %s
+// RUN: %clang -target riscv64-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv64izkt1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZKT-EXT %s
+// CHECK-ZKT-NOT: __riscv_k
+// CHECK-ZKT-EXT: __riscv_zkt 100000
+
+// RUN: %clang -target riscv32-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv32izk1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZK-EXT %s
+// RUN: %clang -target riscv64-unknown-linux-gnu -menable-experimental-extensions \
+// RUN: -march=rv64izk1p0 -x c -E -dM %s \
+// RUN: -o - | FileCheck --check-prefix=CHECK-ZK-EXT %s
+// CHECK-ZK-NOT: __riscv_k
+// CHECK-ZK-EXT: __riscv_zk 100000
+
+// RUN: %clang -target riscv32-unknown-linux-gnu -menable-experimental-extensions \
 // RUN: -march=rv32iv0p10 -x c -E -dM %s \
 // RUN: -o - | FileCheck --check-prefix=CHECK-V-EXT %s
 // RUN: %clang -target riscv64-unknown-linux-gnu -menable-experimental-extensions \
 // RUN: -march=rv64iv0p10 -x c -E -dM %s \
 // RUN: -o - | FileCheck --check-prefix=CHECK-V-EXT %s
 // RUN: %clang -target riscv32-unknown-linux-gnu -menable-experimental-extensions \
 // RUN: -march=rv32izvlsseg0p10 -x c -E -dM %s \
 // RUN: -o - | FileCheck --check-prefix=CHECK-V-EXT %s