[AArch64] Add +pauth archictecture option, allowing the v8.3a pointer authentication extension.

[MarkMurrayARM]

Fixed the parent commit.

[chill]

LGTM

[fhahn]

FWIW I think it would be good to have a bit more details in the description for changes such as this, like a link to the public docs for the extension.

[chill]

In D94083#2486800, @fhahn wrote:

FWIW I think it would be good to have a bit more details in the description for changes such as this, like a link to the public docs for the extension.

I'm sorry, I assumed this information was shared when the extension itself ((this patch just adds the command line options),
was added to LLVM three years ago in https://reviews.llvm.org/D36517

The public documentation is in the Armv8-A ARM at https://developer.arm.com/documentation/ddi0487/latest
in section "D5.1.5 Pointer authentication in AArch64 state"

clang/lib/Basic/Targets/AArch64.h

class LLVM_LIBRARY_VISIBILITY AArch64TargetInfo : public TargetInfo {
   bool HasCRC;
   bool HasCrypto;
   bool HasUnaligned;
   bool HasFullFP16;
   bool HasDotProd;
   bool HasFP16FML;
   bool HasMTE;
   bool HasTME;
+  bool HasPAuth;
   bool HasLS64;
   bool HasMatMul;
   bool HasSVE2;
   bool HasSVE2AES;
   bool HasSVE2SHA3;
   bool HasSVE2SM4;
   bool HasSVE2BitPerm;
   bool HasMatmulFP64;

clang/lib/Basic/Targets/AArch64.cpp

for (const auto &Feature : Features) {
     if (Feature == "+dotprod")
       HasDotProd = true;
     if (Feature == "+fp16fml")
       HasFP16FML = true;
     if (Feature == "+mte")
       HasMTE = true;
     if (Feature == "+tme")
       HasTME = true;
+    if (Feature == "+pauth")
+      HasPAuth = true;
     if (Feature == "+i8mm")
       HasMatMul = true;
     if (Feature == "+bf16")
       HasBFloat16 = true;
     if (Feature == "+lse")
       HasLSE = true;
     if (Feature == "+ls64")
       HasLS64 = true;

llvm/include/llvm/Support/AArch64TargetParser.h

enum ArchExtKind : uint64_t {
   AEK_SVE2BITPERM = 1 << 27,
   AEK_TME =         1 << 28,
   AEK_BF16 =        1 << 29,
   AEK_I8MM =        1 << 30,
   AEK_F32MM =       1ULL << 31,
   AEK_F64MM =       1ULL << 32,
   AEK_LS64 =        1ULL << 33,
   AEK_BRBE =        1ULL << 34,
+  AEK_PAUTH =       1ULL << 35,
 };
 
 enum class ArchKind {
 #define AARCH64_ARCH(NAME, ID, CPU_ATTR, SUB_ARCH, ARCH_ATTR, ARCH_FPU, ARCH_BASE_EXT) ID,
 #include "AArch64TargetParser.def"
 };
 
 const ARM::ArchNames<ArchKind> AArch64ARCHNames[] = {

llvm/include/llvm/Support/AArch64TargetParser.def

 AARCH64_ARCH_EXT_NAME("predres",      AArch64::AEK_PREDRES,     "+predres", "-predres")
 AARCH64_ARCH_EXT_NAME("bf16",         AArch64::AEK_BF16,        "+bf16",  "-bf16")
 AARCH64_ARCH_EXT_NAME("i8mm",         AArch64::AEK_I8MM,        "+i8mm",  "-i8mm")
 AARCH64_ARCH_EXT_NAME("f32mm",        AArch64::AEK_F32MM,       "+f32mm", "-f32mm")
 AARCH64_ARCH_EXT_NAME("f64mm",        AArch64::AEK_F64MM,       "+f64mm", "-f64mm")
 AARCH64_ARCH_EXT_NAME("tme",          AArch64::AEK_TME,         "+tme",   "-tme")
 AARCH64_ARCH_EXT_NAME("ls64",         AArch64::AEK_LS64,        "+ls64",  "-ls64")
 AARCH64_ARCH_EXT_NAME("brbe",         AArch64::AEK_BRBE,        "+brbe",  "-brbe")
+AARCH64_ARCH_EXT_NAME("pauth",        AArch64::AEK_PAUTH,       "+pauth", "-pauth")
 #undef AARCH64_ARCH_EXT_NAME
 
 #ifndef AARCH64_CPU_NAME
 #define AARCH64_CPU_NAME(NAME, ID, DEFAULT_FPU, IS_DEFAULT, DEFAULT_EXT)
 #endif
 AARCH64_CPU_NAME("cortex-a34", ARMV8A, FK_CRYPTO_NEON_FP_ARMV8, false,
                  (AArch64::AEK_CRC))
 AARCH64_CPU_NAME("cortex-a35", ARMV8A, FK_CRYPTO_NEON_FP_ARMV8, false,

llvm/lib/Support/AArch64TargetParser.cpp

bool AArch64::getExtensionFeatures(uint64_t Extensions,
   if (Extensions & AEK_SVE2SHA3)
     Features.push_back("+sve2-sha3");
   if (Extensions & AEK_SVE2BITPERM)
     Features.push_back("+sve2-bitperm");
   if (Extensions & AEK_RCPC)
     Features.push_back("+rcpc");
   if (Extensions & AEK_BRBE)
     Features.push_back("+brbe");
+  if (Extensions & AEK_PAUTH)
+    Features.push_back("+pauth");
 
   return true;
 }
 
 bool AArch64::getArchFeatures(AArch64::ArchKind AK,
                               std::vector<StringRef> &Features) {
   if (AK == ArchKind::ARMV8_1A)
     Features.push_back("+v8.1a");

llvm/lib/Target/AArch64/AArch64.td

 def FeatureUseRSqrt : SubtargetFeature<
     "use-reciprocal-square-root", "UseRSqrt", "true",
     "Use the reciprocal square root approximation">;
 
 def FeatureDotProd : SubtargetFeature<
     "dotprod", "HasDotProd", "true",
     "Enable dot product support">;
 
-def FeaturePA : SubtargetFeature<
-    "pa", "HasPA", "true",
+def FeaturePAuth : SubtargetFeature<
+    "pauth", "HasPAuth", "true",
     "Enable v8.3-A Pointer Authentication extension">;
 
 def FeatureJS : SubtargetFeature<
     "jsconv", "HasJS", "true",
     "Enable v8.3-A JavaScript FP conversion instructions",
     [FeatureFPARMv8]>;
 
 def FeatureCCIDX : SubtargetFeature<
def HasV8_1aOps : SubtargetFeature<"v8.1a", "HasV8_1aOps", "true",
   "Support ARM v8.1a instructions", [FeatureCRC, FeatureLSE, FeatureRDM,
   FeaturePAN, FeatureLOR, FeatureVH]>;
 
 def HasV8_2aOps : SubtargetFeature<"v8.2a", "HasV8_2aOps", "true",
   "Support ARM v8.2a instructions", [HasV8_1aOps, FeaturePsUAO,
   FeaturePAN_RWV, FeatureRAS, FeatureCCPP]>;
 
 def HasV8_3aOps : SubtargetFeature<"v8.3a", "HasV8_3aOps", "true",
-  "Support ARM v8.3a instructions", [HasV8_2aOps, FeatureRCPC, FeaturePA,
+  "Support ARM v8.3a instructions", [HasV8_2aOps, FeatureRCPC, FeaturePAuth,
   FeatureJS, FeatureCCIDX, FeatureComplxNum]>;
 
 def HasV8_4aOps : SubtargetFeature<"v8.4a", "HasV8_4aOps", "true",
   "Support ARM v8.4a instructions", [HasV8_3aOps, FeatureDotProd,
   FeatureNV, FeatureMPAM, FeatureDIT,
   FeatureTRACEV8_4, FeatureAM, FeatureSEL2, FeaturePMU, FeatureTLB_RMI,
   FeatureFMI, FeatureRCPC_IMMO]>;
 
def HasV8_0rOps : SubtargetFeature<
   "v8r", "HasV8_0rOps", "true", "Support ARM v8r instructions",
   [//v8.1
   FeatureCRC, FeaturePAN, FeatureRDM, FeatureLSE, FeatureCONTEXTIDREL2,
   //v8.2
   FeaturePerfMon, FeatureRAS, FeaturePsUAO, FeatureSM4,
   FeatureSHA3, FeatureCCPP, FeatureFullFP16, FeaturePAN_RWV,
   //v8.3
   FeatureComplxNum, FeatureCCIDX, FeatureJS,
-  FeaturePA, FeatureRCPC,
+  FeaturePAuth, FeatureRCPC,
   //v8.4
   FeatureDotProd, FeatureFP16FML, FeatureTRACEV8_4,
   FeatureTLB_RMI, FeatureFMI, FeatureDIT, FeatureSEL2, FeatureRCPC_IMMO,
   //v8.5
   FeatureSSBS, FeaturePredRes, FeatureSB, FeatureSpecRestrict]>;
 
 //===----------------------------------------------------------------------===//
 // Register File Description

 class AArch64Unsupported { list<Predicate> F; }
 
 def SVEUnsupported : AArch64Unsupported {
   let F = [HasSVE, HasSVE2, HasSVE2AES, HasSVE2SM4, HasSVE2SHA3,
            HasSVE2BitPerm];
 }
 
 def PAUnsupported : AArch64Unsupported {
-  let F = [HasPA];
+  let F = [HasPAuth];
 }
 
 include "AArch64SchedA53.td"
 include "AArch64SchedA55.td"
 include "AArch64SchedA57.td"
 include "AArch64SchedCyclone.td"
 include "AArch64SchedFalkor.td"
 include "AArch64SchedKryo.td"
def ProcA78C : SubtargetFeature<"cortex-a78c", "ARMProcFamily",
                                 FeatureCrypto,
                                 FeatureDotProd,
                                 FeatureFMI,
                                 FeatureFP16FML,
                                 FeatureFPARMv8,
                                 FeatureFullFP16,
                                 FeatureFuseAES,
                                 FeatureNEON,
-                                FeaturePA,
+                                FeaturePAuth,
                                 FeaturePerfMon,
                                 FeaturePostRAScheduler,
                                 FeatureRCPC,
                                 FeatureSPE,
                                 FeatureSSBS]>;
 
 def ProcR82 : SubtargetFeature<"cortex-r82", "ARMProcFamily",
                                "CortexR82",
def ProcThunderX3T110  : SubtargetFeature<"thunderx3t110", "ARMProcFamily",
                                           FeatureCRC,
                                           FeatureCrypto,
                                           FeatureFPARMv8,
                                           FeatureArithmeticBccFusion,
                                           FeatureNEON,
                                           FeaturePostRAScheduler,
                                           FeaturePredictableSelectIsExpensive,
                                           FeatureLSE,
-                                          FeaturePA,
+                                          FeaturePAuth,
                                           FeatureUseAA,
                                           FeatureBalanceFPOps,
                                           FeaturePerfMon,
                                           FeatureStrictAlign,
                                           HasV8_3aOps]>;
 
 def ProcThunderX : SubtargetFeature<"thunderx", "ARMProcFamily", "ThunderX",
                                     "Cavium ThunderX processors", [

llvm/lib/Target/AArch64/AArch64FrameLowering.cpp

static void InsertReturnAddressAuth(MachineFunction &MF,
   DebugLoc DL;
   if (MBBI != MBB.end())
     DL = MBBI->getDebugLoc();
 
   // The AUTIASP instruction assembles to a hint instruction before v8.3a so
   // this instruction can safely used for any v8a architecture.
   // From v8.3a onwards there are optimised authenticate LR and return
   // instructions, namely RETA{A,B}, that can be used instead.
-  if (Subtarget.hasV8_3aOps() && MBBI != MBB.end() &&
+  if (Subtarget.hasPAuth() && MBBI != MBB.end() &&
       MBBI->getOpcode() == AArch64::RET_ReallyLR) {
     BuildMI(MBB, MBBI, DL,
             TII->get(MFI.shouldSignWithBKey() ? AArch64::RETAB : AArch64::RETAA))
         .copyImplicitOps(*MBBI);
     MBB.erase(MBBI);
   } else {
     BuildMI(
         MBB, MBBI, DL,

llvm/lib/Target/AArch64/AArch64ISelLowering.cpp

if (Depth) {
     ReturnAddress = DAG.getCopyFromReg(DAG.getEntryNode(), DL, Reg, VT);
   }
 
   // The XPACLRI instruction assembles to a hint-space instruction before
   // Armv8.3-A therefore this instruction can be safely used for any pre
   // Armv8.3-A architectures. On Armv8.3-A and onwards XPACI is available so use
   // that instead.
   SDNode *St;
-  if (Subtarget->hasV8_3aOps()) {
+  if (Subtarget->hasPAuth()) {
     St = DAG.getMachineNode(AArch64::XPACI, DL, VT, ReturnAddress);
   } else {
     // XPACLRI operates on LR therefore we must move the operand accordingly.
     SDValue Chain =
         DAG.getCopyToReg(DAG.getEntryNode(), DL, AArch64::LR, ReturnAddress);
     St = DAG.getMachineNode(AArch64::XPACLRI, DL, VT, Chain);
   }
   return SDValue(St, 0);

llvm/lib/Target/AArch64/AArch64InstrInfo.cpp

if (ShouldSignReturnAddr) {
     unsigned CFIIndex =
         MF.addFrameInst(MCCFIInstruction::createNegateRAState(nullptr));
     BuildMI(MBB, MBBPAC, DebugLoc(), TII->get(AArch64::CFI_INSTRUCTION))
         .addCFIIndex(CFIIndex)
         .setMIFlags(MachineInstr::FrameSetup);
 
     // If v8.3a features are available we can replace a RET instruction by
     // RETAA or RETAB and omit the AUT instructions
-    if (Subtarget.hasV8_3aOps() && MBBAUT != MBB.end() &&
+    if (Subtarget.hasPAuth() && MBBAUT != MBB.end() &&
         MBBAUT->getOpcode() == AArch64::RET) {
       BuildMI(MBB, MBBAUT, DL,
               TII->get(ShouldSignReturnAddrWithAKey ? AArch64::RETAA
                                                     : AArch64::RETAB))
           .copyImplicitOps(*MBBAUT);
       MBB.erase(MBBAUT);
     } else {
       BuildMI(MBB, MBBAUT, DL,

llvm/lib/Target/AArch64/AArch64InstrInfo.td

 def HasV8_7a         : Predicate<"Subtarget->hasV8_7aOps()">,
                                  AssemblerPredicate<(all_of HasV8_7aOps), "armv8.7a">;
 def HasVH            : Predicate<"Subtarget->hasVH()">,
                        AssemblerPredicate<(all_of FeatureVH), "vh">;
 
 def HasLOR           : Predicate<"Subtarget->hasLOR()">,
                        AssemblerPredicate<(all_of FeatureLOR), "lor">;
 
-def HasPA            : Predicate<"Subtarget->hasPA()">,
-                       AssemblerPredicate<(all_of FeaturePA), "pa">;
+def HasPAuth         : Predicate<"Subtarget->hasPAuth()">,
+                       AssemblerPredicate<(all_of FeaturePAuth), "pauth">;
 
 def HasJS            : Predicate<"Subtarget->hasJS()">,
                        AssemblerPredicate<(all_of FeatureJS), "jsconv">;
 
 def HasCCIDX         : Predicate<"Subtarget->hasCCIDX()">,
                        AssemblerPredicate<(all_of FeatureCCIDX), "ccidx">;
 
 def HasComplxNum      : Predicate<"Subtarget->hasComplxNum()">,

 def : InstAlias<"autibsp", (AUTIBSP), 0>;
 def : InstAlias<"pacia1716", (PACIA1716), 0>;
 def : InstAlias<"pacib1716", (PACIB1716), 0>;
 def : InstAlias<"autia1716", (AUTIA1716), 0>;
 def : InstAlias<"autib1716", (AUTIB1716), 0>;
 def : InstAlias<"xpaclri", (XPACLRI), 0>;
 
 // These pointer authentication instructions require armv8.3a
-let Predicates = [HasPA] in {
+let Predicates = [HasPAuth] in {
 
   // When PA is enabled, a better mnemonic should be emitted.
   def : InstAlias<"paciaz", (PACIAZ), 1>;
   def : InstAlias<"pacibz", (PACIBZ), 1>;
   def : InstAlias<"autiaz", (AUTIAZ), 1>;
   def : InstAlias<"autibz", (AUTIBZ), 1>;
   def : InstAlias<"paciasp", (PACIASP), 1>;
   def : InstAlias<"pacibsp", (PACIBSP), 1>;

llvm/lib/Target/AArch64/AArch64Subtarget.h

protected:
 
   // Armv8.2 Crypto extensions
   bool HasSM4 = false;
   bool HasSHA3 = false;
   bool HasSHA2 = false;
   bool HasAES = false;
 
   // ARMv8.3 extensions
-  bool HasPA = false;
+  bool HasPAuth = false;
   bool HasJS = false;
   bool HasCCIDX = false;
   bool HasComplxNum = false;
 
   // ARMv8.4 extensions
   bool HasNV = false;
   bool HasMPAM = false;
   bool HasDIT = false;
protected:
   bool HasSVE2SM4 = false;
   bool HasSVE2SHA3 = false;
   bool HasSVE2BitPerm = false;
 
   // Future architecture extensions.
   bool HasETE = false;
   bool HasTRBE = false;
   bool HasBRBE = false;
+  bool HasPAUTH = false;
   bool HasSPE_EEF = false;
 
   // HasZeroCycleRegMove - Has zero-cycle register mov instructions.
   bool HasZeroCycleRegMove = false;
 
   // HasZeroCycleZeroing - Has zero-cycle zeroing instructions.
   bool HasZeroCycleZeroing = false;
   bool HasZeroCycleZeroingGP = false;
public:
   bool hasSSBS() const { return HasSSBS; }
   bool hasSB() const { return HasSB; }
   bool hasPredRes() const { return HasPredRes; }
   bool hasCCDP() const { return HasCCDP; }
   bool hasBTI() const { return HasBTI; }
   bool hasRandGen() const { return HasRandGen; }
   bool hasMTE() const { return HasMTE; }
   bool hasTME() const { return HasTME; }
+  bool hasPAUTH() const { return HasPAUTH; }
   // Arm SVE2 extensions
   bool hasSVE2AES() const { return HasSVE2AES; }
   bool hasSVE2SM4() const { return HasSVE2SM4; }
   bool hasSVE2SHA3() const { return HasSVE2SHA3; }
   bool hasSVE2BitPerm() const { return HasSVE2BitPerm; }
   bool hasMatMulInt8() const { return HasMatMulInt8; }
   bool hasMatMulFP32() const { return HasMatMulFP32; }
   bool hasMatMulFP64() const { return HasMatMulFP64; }
public:
   bool hasVH() const { return HasVH; }
   bool hasPAN() const { return HasPAN; }
   bool hasLOR() const { return HasLOR; }
 
   bool hasPsUAO() const { return HasPsUAO; }
   bool hasPAN_RWV() const { return HasPAN_RWV; }
   bool hasCCPP() const { return HasCCPP; }
 
-  bool hasPA() const { return HasPA; }
+  bool hasPAuth() const { return HasPAuth; }
   bool hasJS() const { return HasJS; }
   bool hasCCIDX() const { return HasCCIDX; }
   bool hasComplxNum() const { return HasComplxNum; }
 
   bool hasNV() const { return HasNV; }
   bool hasMPAM() const { return HasMPAM; }
   bool hasDIT() const { return HasDIT; }
   bool hasTRACEV8_4() const { return HasTRACEV8_4; }

llvm/lib/Target/AArch64/AArch64SystemOperands.td

 def : RWSysReg<"ERXMISC1_EL1",  0b11, 0b000, 0b0101, 0b0101, 0b001>;
 def : RWSysReg<"DISR_EL1",      0b11, 0b000, 0b1100, 0b0001, 0b001>;
 def : RWSysReg<"VDISR_EL2",     0b11, 0b100, 0b1100, 0b0001, 0b001>;
 def : RWSysReg<"VSESR_EL2",     0b11, 0b100, 0b0101, 0b0010, 0b011>;
 }
 
 // v8.3a "Pointer authentication extension" registers
 //                              Op0    Op1     CRn     CRm    Op2
-let Requires = [{ {AArch64::FeaturePA} }] in {
+let Requires = [{ {AArch64::FeaturePAuth} }] in {
 def : RWSysReg<"APIAKeyLo_EL1", 0b11, 0b000, 0b0010, 0b0001, 0b000>;
 def : RWSysReg<"APIAKeyHi_EL1", 0b11, 0b000, 0b0010, 0b0001, 0b001>;
 def : RWSysReg<"APIBKeyLo_EL1", 0b11, 0b000, 0b0010, 0b0001, 0b010>;
 def : RWSysReg<"APIBKeyHi_EL1", 0b11, 0b000, 0b0010, 0b0001, 0b011>;
 def : RWSysReg<"APDAKeyLo_EL1", 0b11, 0b000, 0b0010, 0b0010, 0b000>;
 def : RWSysReg<"APDAKeyHi_EL1", 0b11, 0b000, 0b0010, 0b0010, 0b001>;
 def : RWSysReg<"APDBKeyLo_EL1", 0b11, 0b000, 0b0010, 0b0010, 0b010>;
 def : RWSysReg<"APDBKeyHi_EL1", 0b11, 0b000, 0b0010, 0b0010, 0b011>;

llvm/lib/Target/AArch64/GISel/AArch64InstructionSelector.cpp

if (Depth == 0 && IntrinID == Intrinsic::returnaddress) {
       if (!MFReturnAddr) {
         // Insert the copy from LR/X30 into the entry block, before it can be
         // clobbered by anything.
         MFI.setReturnAddressIsTaken(true);
         MFReturnAddr = getFunctionLiveInPhysReg(MF, TII, AArch64::LR,
                                                 AArch64::GPR64RegClass);
       }
 
-      if (STI.hasV8_3aOps()) {
+      if (STI.hasPAuth()) {
         MIRBuilder.buildInstr(AArch64::XPACI, {DstReg}, {MFReturnAddr});
       } else {
         MIRBuilder.buildCopy({Register(AArch64::LR)}, {MFReturnAddr});
         MIRBuilder.buildInstr(AArch64::XPACLRI);
         MIRBuilder.buildCopy({DstReg}, {Register(AArch64::LR)});
       }
 
       I.eraseFromParent();
while (Depth--) {
       FrameAddr = NextFrame;
     }
 
     if (IntrinID == Intrinsic::frameaddress)
       MIRBuilder.buildCopy({DstReg}, {FrameAddr});
     else {
       MFI.setReturnAddressIsTaken(true);
 
-      if (STI.hasV8_3aOps()) {
+      if (STI.hasPAuth()) {
         Register TmpReg = MRI.createVirtualRegister(&AArch64::GPR64RegClass);
         MIRBuilder.buildInstr(AArch64::LDRXui, {TmpReg}, {FrameAddr}).addImm(1);
         MIRBuilder.buildInstr(AArch64::XPACI, {DstReg}, {TmpReg});
       } else {
         MIRBuilder.buildInstr(AArch64::LDRXui, {Register(AArch64::LR)}, {FrameAddr}).addImm(1);
         MIRBuilder.buildInstr(AArch64::XPACLRI);
         MIRBuilder.buildCopy({DstReg}, {Register(AArch64::LR)});
       }

llvm/unittests/Support/TargetParserTest.cpp

const char *ArchExt[][4] = {{"crc", "nocrc", "+crc", "-crc"},
                                "-sve2-sha3"},
                               {"sve2-bitperm", "nosve2-bitperm",
                                "+sve2-bitperm", "-sve2-bitperm"},
                               {"dotprod", "nodotprod", "+dotprod", "-dotprod"},
                               {"rcpc", "norcpc", "+rcpc", "-rcpc" },
                               {"rng", "norng", "+rand", "-rand"},
                               {"memtag", "nomemtag", "+mte", "-mte"},
                               {"tme", "notme", "+tme", "-tme"},
+                              {"pauth", "nopauth", "+pauth", "-pauth"},
                               {"ssbs", "nossbs", "+ssbs", "-ssbs"},
                               {"sb", "nosb", "+sb", "-sb"},
                               {"predres", "nopredres", "+predres", "-predres"},
                               {"i8mm", "noi8mm", "+i8mm", "-i8mm"},
                               {"f32mm", "nof32mm", "+f32mm", "-f32mm"},
                               {"f64mm", "nof64mm", "+f64mm", "-f64mm"},
 };