From dc198344ce15a7d25c5926a284f28dbc1ba30d97 Mon Sep 17 00:00:00 2001 From: Sjoerd Meijer Date: Thu, 26 Jul 2018 07:13:59 +0000 Subject: [PATCH] [AArch64] Armv8.2-A: add the crypto extensions This adds MC support for the crypto instructions that were made optional extensions in Armv8.2-A (AArch64 only). Differential Revision: https://reviews.llvm.org/D49370 llvm-svn: 338010 --- llvm/include/llvm/Support/TargetParser.h | 1 + llvm/lib/Support/TargetParser.cpp | 11 +++ llvm/lib/Target/AArch64/AArch64InstrFormats.td | 101 ++++++++++++++++++++- llvm/lib/Target/AArch64/AArch64InstrInfo.td | 38 +++++++- .../Target/AArch64/AsmParser/AArch64AsmParser.cpp | 58 +++++++++++- llvm/test/MC/AArch64/arm64-diagno-predicate.s | 2 +- llvm/test/MC/AArch64/armv8.2a-crypto-error.s | 31 +++++++ llvm/test/MC/AArch64/armv8.2a-crypto.s | 64 +++++++++++++ llvm/test/MC/AArch64/directive-cpu-err.s | 26 +++++- llvm/test/MC/AArch64/directive-cpu.s | 20 +++- llvm/test/MC/AArch64/neon-crypto.s | 15 ++- llvm/test/MC/AArch64/nofp-crypto-diagnostic.s | 2 +- .../MC/Disassembler/AArch64/armv8.2a-crypto.txt | 93 +++++++++++++++++++ 13 files changed, 444 insertions(+), 18 deletions(-) create mode 100644 llvm/test/MC/AArch64/armv8.2a-crypto-error.s create mode 100644 llvm/test/MC/AArch64/armv8.2a-crypto.s create mode 100644 llvm/test/MC/Disassembler/AArch64/armv8.2a-crypto.txt diff --git a/llvm/include/llvm/Support/TargetParser.h b/llvm/include/llvm/Support/TargetParser.h index 716a62b..08ad42d 100644 --- a/llvm/include/llvm/Support/TargetParser.h +++ b/llvm/include/llvm/Support/TargetParser.h @@ -206,6 +206,7 @@ unsigned checkArchVersion(StringRef Arch); unsigned getDefaultFPU(StringRef CPU, ArchKind AK); unsigned getDefaultExtensions(StringRef CPU, ArchKind AK); StringRef getDefaultCPU(StringRef Arch); +AArch64::ArchKind getCPUArchKind(StringRef CPU); // Parser unsigned parseFPU(StringRef FPU); diff --git a/llvm/lib/Support/TargetParser.cpp b/llvm/lib/Support/TargetParser.cpp index bb2aecd..2c167a4 100644 --- a/llvm/lib/Support/TargetParser.cpp +++ b/llvm/lib/Support/TargetParser.cpp @@ -433,6 +433,17 @@ unsigned llvm::AArch64::getDefaultExtensions(StringRef CPU, ArchKind AK) { .Default(AArch64::AEK_INVALID); } +AArch64::ArchKind llvm::AArch64::getCPUArchKind(StringRef CPU) { + if (CPU == "generic") + return AArch64::ArchKind::ARMV8A; + + return StringSwitch(CPU) +#define AARCH64_CPU_NAME(NAME, ID, DEFAULT_FPU, IS_DEFAULT, DEFAULT_EXT) \ + .Case(NAME, AArch64::ArchKind:: ID) +#include "llvm/Support/AArch64TargetParser.def" + .Default(AArch64::ArchKind::INVALID); +} + bool llvm::AArch64::getExtensionFeatures(unsigned Extensions, std::vector &Features) { diff --git a/llvm/lib/Target/AArch64/AArch64InstrFormats.td b/llvm/lib/Target/AArch64/AArch64InstrFormats.td index 1ba2f38..1060c64 100644 --- a/llvm/lib/Target/AArch64/AArch64InstrFormats.td +++ b/llvm/lib/Target/AArch64/AArch64InstrFormats.td @@ -5720,7 +5720,7 @@ multiclass SIMDDifferentThreeVectorBD opc, string asm, def v16i8 : BaseSIMDDifferentThreeVector; - let Predicates = [HasCrypto] in { + let Predicates = [HasAES] in { def v1i64 : BaseSIMDDifferentThreeVector; @@ -9920,7 +9920,6 @@ multiclass SIMDIndexedTiedComplexHSD opc, string asm, dag outs, dag ins, string cstr, list pat> @@ -10010,7 +10009,103 @@ class SHATiedInstVV opc, string asm, Intrinsic OpNode> class SHAInstSS opc, string asm, Intrinsic OpNode> : SHA2OpInst; -} // end of 'let Predicates = [HasCrypto]' + +// Armv8.2-A Crypto extensions +class BaseCryptoV82 pattern> + : I , Sched<[WriteV]> { + bits<5> Vd; + bits<5> Vn; + let Inst{31-25} = 0b1100111; + let Inst{9-5} = Vn; + let Inst{4-0} = Vd; +} + +class CryptoRRTiedop0, bits<2>op1, string asm, string asmops> + : BaseCryptoV82<(outs V128:$Vd), (ins V128:$Vn, V128:$Vm), asm, asmops, + "$Vm = $Vd", []> { + let Inst{31-25} = 0b1100111; + let Inst{24-21} = 0b0110; + let Inst{20-15} = 0b000001; + let Inst{14} = op0; + let Inst{13-12} = 0b00; + let Inst{11-10} = op1; +} +class CryptoRRTied_2Dop0, bits<2>op1, string asm> + : CryptoRRTied; +class CryptoRRTied_4Sop0, bits<2>op1, string asm> + : CryptoRRTied; + +class CryptoRRR op0, bits<2>op1, dag oops, dag iops, string asm, + string asmops, string cst> + : BaseCryptoV82 { + bits<5> Vm; + let Inst{24-21} = 0b0011; + let Inst{20-16} = Vm; + let Inst{15} = 0b1; + let Inst{14} = op0; + let Inst{13-12} = 0b00; + let Inst{11-10} = op1; +} +class CryptoRRR_2D op0, bits<2>op1, string asm> + : CryptoRRR; +class CryptoRRRTied_2D op0, bits<2>op1, string asm> + : CryptoRRR; +class CryptoRRR_4S op0, bits<2>op1, string asm> + : CryptoRRR; +class CryptoRRRTied_4S op0, bits<2>op1, string asm> + : CryptoRRR; +class CryptoRRRTied op0, bits<2>op1, string asm> + : CryptoRRR; + +class CryptoRRRRop0, string asm, string asmops> + : BaseCryptoV82<(outs V128:$Vd), (ins V128:$Vn, V128:$Vm, V128:$Va), asm, + asmops, "", []> { + bits<5> Vm; + bits<5> Va; + let Inst{24-23} = 0b00; + let Inst{22-21} = op0; + let Inst{20-16} = Vm; + let Inst{15} = 0b0; + let Inst{14-10} = Va; +} +class CryptoRRRR_16Bop0, string asm> + : CryptoRRRR { +} +class CryptoRRRR_4Sop0, string asm> + : CryptoRRRR { +} + +class CryptoRRRi6 + : BaseCryptoV82<(outs V128:$Vd), (ins V128:$Vn, V128:$Vm, uimm6:$imm), asm, + "{\t$Vd.2d, $Vn.2d, $Vm.2d, $imm}", "", []> { + bits<6> imm; + bits<5> Vm; + let Inst{24-21} = 0b0100; + let Inst{20-16} = Vm; + let Inst{15-10} = imm; + let Inst{9-5} = Vn; + let Inst{4-0} = Vd; +} + +class CryptoRRRi2Tiedop0, bits<2>op1, string asm> + : BaseCryptoV82<(outs V128:$Vdst), + (ins V128:$Vd, V128:$Vn, V128:$Vm, VectorIndexS:$imm), + asm, "{\t$Vd.4s, $Vn.4s, $Vm.s$imm}", "$Vd = $Vdst", []> { + bits<2> imm; + bits<5> Vm; + let Inst{24-21} = 0b0010; + let Inst{20-16} = Vm; + let Inst{15} = 0b1; + let Inst{14} = op0; + let Inst{13-12} = imm; + let Inst{11-10} = op1; +} //---------------------------------------------------------------------------- // v8.1 atomic instructions extension: diff --git a/llvm/lib/Target/AArch64/AArch64InstrInfo.td b/llvm/lib/Target/AArch64/AArch64InstrInfo.td index 6ea7b01..62c9599 100644 --- a/llvm/lib/Target/AArch64/AArch64InstrInfo.td +++ b/llvm/lib/Target/AArch64/AArch64InstrInfo.td @@ -28,6 +28,14 @@ def HasNEON : Predicate<"Subtarget->hasNEON()">, AssemblerPredicate<"FeatureNEON", "neon">; def HasCrypto : Predicate<"Subtarget->hasCrypto()">, AssemblerPredicate<"FeatureCrypto", "crypto">; +def HasSM4 : Predicate<"Subtarget->hasSM4()">, + AssemblerPredicate<"FeatureSM4", "sm4">; +def HasSHA3 : Predicate<"Subtarget->hasSHA3()">, + AssemblerPredicate<"FeatureSHA3", "sha3">; +def HasSHA2 : Predicate<"Subtarget->hasSHA2()">, + AssemblerPredicate<"FeatureSHA2", "sha2">; +def HasAES : Predicate<"Subtarget->hasAES()">, + AssemblerPredicate<"FeatureAES", "aes">; def HasDotProd : Predicate<"Subtarget->hasDotProd()">, AssemblerPredicate<"FeatureDotProd", "dotprod">; def HasCRC : Predicate<"Subtarget->hasCRC()">, @@ -467,6 +475,30 @@ defm SDOTlane : SIMDThreeSameVectorDotIndex<0, "sdot", int_aarch64_neon_sdot>; defm UDOTlane : SIMDThreeSameVectorDotIndex<1, "udot", int_aarch64_neon_udot>; } +// Armv8.2-A Crypto extensions +let Predicates = [HasSHA3] in { +def SHA512H : CryptoRRRTied<0b0, 0b00, "sha512h">; +def SHA512H2 : CryptoRRRTied<0b0, 0b01, "sha512h2">; +def SHA512SU0 : CryptoRRTied_2D<0b0, 0b00, "sha512su0">; +def SHA512SU1 : CryptoRRRTied_2D<0b0, 0b10, "sha512su1">; +def RAX1 : CryptoRRR_2D<0b0,0b11, "rax1">; +def EOR3 : CryptoRRRR_16B<0b00, "eor3">; +def BCAX : CryptoRRRR_16B<0b01, "bcax">; +def XAR : CryptoRRRi6<"xar">; +} // HasSHA3 + +let Predicates = [HasSM4] in { +def SM3TT1A : CryptoRRRi2Tied<0b0, 0b00, "sm3tt1a">; +def SM3TT1B : CryptoRRRi2Tied<0b0, 0b01, "sm3tt1b">; +def SM3TT2A : CryptoRRRi2Tied<0b0, 0b10, "sm3tt2a">; +def SM3TT2B : CryptoRRRi2Tied<0b0, 0b11, "sm3tt2b">; +def SM3SS1 : CryptoRRRR_4S<0b10, "sm3ss1">; +def SM3PARTW1 : CryptoRRRTied_4S<0b1, 0b00, "sm3partw1">; +def SM3PARTW2 : CryptoRRRTied_4S<0b1, 0b01, "sm3partw2">; +def SM4ENCKEY : CryptoRRR_4S<0b1, 0b10, "sm4ekey">; +def SM4E : CryptoRRTied_4S<0b0, 0b01, "sm4e">; +} // HasSM4 + let Predicates = [HasRCPC] in { // v8.3 Release Consistent Processor Consistent support, optional in v8.2. def LDAPRB : RCPCLoad<0b00, "ldaprb", GPR32>; @@ -555,7 +587,7 @@ let Predicates = [HasV8_3a] in { let Inst{31} = 0; } -} // HasV8_3A +} // HasV8_3a // v8.4 Flag manipulation instructions let Predicates = [HasV8_4a] in { @@ -5606,10 +5638,12 @@ defm ST4 : SIMDLdSt4SingleAliases<"st4">; // Crypto extensions //---------------------------------------------------------------------------- +let Predicates = [HasAES] in { def AESErr : AESTiedInst<0b0100, "aese", int_aarch64_crypto_aese>; def AESDrr : AESTiedInst<0b0101, "aesd", int_aarch64_crypto_aesd>; def AESMCrr : AESInst< 0b0110, "aesmc", int_aarch64_crypto_aesmc>; def AESIMCrr : AESInst< 0b0111, "aesimc", int_aarch64_crypto_aesimc>; +} // Pseudo instructions for AESMCrr/AESIMCrr with a register constraint required // for AES fusion on some CPUs. @@ -5636,6 +5670,7 @@ def : Pat<(v16i8 (int_aarch64_crypto_aesimc (v16i8 V128:$src2)))))>, Requires<[HasFuseAES]>; +let Predicates = [HasSHA2] in { def SHA1Crrr : SHATiedInstQSV<0b000, "sha1c", int_aarch64_crypto_sha1c>; def SHA1Prrr : SHATiedInstQSV<0b001, "sha1p", int_aarch64_crypto_sha1p>; def SHA1Mrrr : SHATiedInstQSV<0b010, "sha1m", int_aarch64_crypto_sha1m>; @@ -5647,6 +5682,7 @@ def SHA256SU1rrr :SHATiedInstVVV<0b110, "sha256su1",int_aarch64_crypto_sha256su1 def SHA1Hrr : SHAInstSS< 0b0000, "sha1h", int_aarch64_crypto_sha1h>; def SHA1SU1rr : SHATiedInstVV<0b0001, "sha1su1", int_aarch64_crypto_sha1su1>; def SHA256SU0rr : SHATiedInstVV<0b0010, "sha256su0",int_aarch64_crypto_sha256su0>; +} //---------------------------------------------------------------------------- // Compiler-pseudos diff --git a/llvm/lib/Target/AArch64/AsmParser/AArch64AsmParser.cpp b/llvm/lib/Target/AArch64/AsmParser/AArch64AsmParser.cpp index 9c8f9d1..a51c41d 100644 --- a/llvm/lib/Target/AArch64/AsmParser/AArch64AsmParser.cpp +++ b/llvm/lib/Target/AArch64/AsmParser/AArch64AsmParser.cpp @@ -4749,7 +4749,11 @@ static const struct { const char *Name; const FeatureBitset Features; } ExtensionMap[] = { - { "crc", {AArch64::FeatureCRC} }, + { "crc", {AArch64::FeatureCRC} }, + { "sm4", {AArch64::FeatureSM4} }, + { "sha3", {AArch64::FeatureSHA3} }, + { "sha2", {AArch64::FeatureSHA2} }, + { "aes", {AArch64::FeatureAES} }, { "crypto", {AArch64::FeatureCrypto} }, { "fp", {AArch64::FeatureFPARMv8} }, { "simd", {AArch64::FeatureNEON} }, @@ -4763,6 +4767,54 @@ static const struct { { "profile", {} }, }; +static void ExpandCryptoAEK(AArch64::ArchKind ArchKind, + SmallVector &RequestedExtensions) { + const bool NoCrypto = + (std::find(RequestedExtensions.begin(), RequestedExtensions.end(), + "nocrypto") != std::end(RequestedExtensions)); + const bool Crypto = + (std::find(RequestedExtensions.begin(), RequestedExtensions.end(), + "crypto") != std::end(RequestedExtensions)); + + if (!NoCrypto && Crypto) { + switch (ArchKind) { + default: + // Map 'generic' (and others) to sha2 and aes, because + // that was the traditional meaning of crypto. + case AArch64::ArchKind::ARMV8_1A: + case AArch64::ArchKind::ARMV8_2A: + case AArch64::ArchKind::ARMV8_3A: + RequestedExtensions.push_back("sha2"); + RequestedExtensions.push_back("aes"); + break; + case AArch64::ArchKind::ARMV8_4A: + RequestedExtensions.push_back("sm4"); + RequestedExtensions.push_back("sha3"); + RequestedExtensions.push_back("sha2"); + RequestedExtensions.push_back("aes"); + break; + } + } else if (NoCrypto) { + switch (ArchKind) { + default: + // Map 'generic' (and others) to sha2 and aes, because + // that was the traditional meaning of crypto. + case AArch64::ArchKind::ARMV8_1A: + case AArch64::ArchKind::ARMV8_2A: + case AArch64::ArchKind::ARMV8_3A: + RequestedExtensions.push_back("nosha2"); + RequestedExtensions.push_back("noaes"); + break; + case AArch64::ArchKind::ARMV8_4A: + RequestedExtensions.push_back("nosm4"); + RequestedExtensions.push_back("nosha3"); + RequestedExtensions.push_back("nosha2"); + RequestedExtensions.push_back("noaes"); + break; + } + } +} + /// parseDirectiveArch /// ::= .arch token bool AArch64AsmParser::parseDirectiveArch(SMLoc L) { @@ -4793,6 +4845,8 @@ bool AArch64AsmParser::parseDirectiveArch(SMLoc L) { if (!ExtensionString.empty()) ExtensionString.split(RequestedExtensions, '+'); + ExpandCryptoAEK(ID, RequestedExtensions); + FeatureBitset Features = STI.getFeatureBits(); for (auto Name : RequestedExtensions) { bool EnableFeature = true; @@ -4852,6 +4906,8 @@ bool AArch64AsmParser::parseDirectiveCPU(SMLoc L) { STI.setDefaultFeatures(CPU, ""); CurLoc = incrementLoc(CurLoc, CPU.size()); + ExpandCryptoAEK(llvm::AArch64::getCPUArchKind(CPU), RequestedExtensions); + FeatureBitset Features = STI.getFeatureBits(); for (auto Name : RequestedExtensions) { // Advance source location past '+'. diff --git a/llvm/test/MC/AArch64/arm64-diagno-predicate.s b/llvm/test/MC/AArch64/arm64-diagno-predicate.s index 3b757e8..763f260 100644 --- a/llvm/test/MC/AArch64/arm64-diagno-predicate.s +++ b/llvm/test/MC/AArch64/arm64-diagno-predicate.s @@ -13,7 +13,7 @@ // CHECK-ERROR-NEXT: ^ pmull v0.1q, v1.1d, v2.1d -// CHECK-ERROR: error: instruction requires: crypto +// CHECK-ERROR: error: instruction requires: aes // CHECK-ERROR-NEXT: pmull v0.1q, v1.1d, v2.1d // CHECK-ERROR-NEXT: ^ diff --git a/llvm/test/MC/AArch64/armv8.2a-crypto-error.s b/llvm/test/MC/AArch64/armv8.2a-crypto-error.s new file mode 100644 index 0000000..71115ef --- /dev/null +++ b/llvm/test/MC/AArch64/armv8.2a-crypto-error.s @@ -0,0 +1,31 @@ +// RUN: not llvm-mc -triple aarch64 -mattr=+sm4,+sha3 -show-encoding < %s 2>&1 | FileCheck %s + + xar v26.2d, v21.2d, v27.2d, #-1 + xar v26.2d, v21.2d, v27.2d, #64 + sm3tt1a v20.4s, v23.4s, v21.s[4] + sm3tt1b v20.4s, v23.4s, v21.s[4] + sm3tt2a v20.4s, v23.4s, v21.s[4] + sm3tt2b v20.4s, v23.4s, v21.s[4] + sm3tt2b v20.4s, v23.4s, v21.s[-1] + +// CHECK: error: immediate must be an integer in range [0, 63]. +// CHECK-NEXT: xar v26.2d, v21.2d, v27.2d, #-1 +// CHECK-NEXT: ^ +// CHECK-NEXT: error: immediate must be an integer in range [0, 63]. +// CHECK-NEXT: xar v26.2d, v21.2d, v27.2d, #64 +// CHECK-NEXT: ^ +// CHECK-NEXT: error: vector lane must be an integer in range [0, 3]. +// CHECK-NEXT: sm3tt1a v20.4s, v23.4s, v21.s[4] +// CHECK-NEXT: ^ +// CHECK-NEXT: error: vector lane must be an integer in range [0, 3]. +// CHECK-NEXT: sm3tt1b v20.4s, v23.4s, v21.s[4] +// CHECK-NEXT: ^ +// CHECK-NEXT: error: vector lane must be an integer in range [0, 3]. +// CHECK-NEXT: sm3tt2a v20.4s, v23.4s, v21.s[4] +// CHECK-NEXT: ^ +// CHECK-NEXT: error: vector lane must be an integer in range [0, 3]. +// CHECK-NEXT: sm3tt2b v20.4s, v23.4s, v21.s[4] +// CHECK-NEXT: ^ +// CHECK-NEXT: error: vector lane must be an integer in range [0, 3]. +// CHECK-NEXT: sm3tt2b v20.4s, v23.4s, v21.s[-1] +// CHECK-NEXT: ^ diff --git a/llvm/test/MC/AArch64/armv8.2a-crypto.s b/llvm/test/MC/AArch64/armv8.2a-crypto.s new file mode 100644 index 0000000..6bee41d --- /dev/null +++ b/llvm/test/MC/AArch64/armv8.2a-crypto.s @@ -0,0 +1,64 @@ +// RUN: not llvm-mc -triple aarch64 -mattr=+sha3,-sm4 -show-encoding < %s 2> %t | FileCheck %s --check-prefix=CHECK-SHA +// RUN: FileCheck --check-prefix=CHECK-NO-SM < %t %s + +// RUN: not llvm-mc -triple aarch64 -mattr=+sm4,-sha3 -show-encoding < %s 2> %t | FileCheck %s --check-prefix=CHECK-SM +// RUN: FileCheck --check-prefix=CHECK-NO-SHA < %t %s + +// RUN: not llvm-mc -triple aarch64 -show-encoding < %s 2> %t +// RUN: FileCheck --check-prefix=CHECK-NO-SHA --check-prefix=CHECK-NO-SM < %t %s + + sha512h q0, q1, v2.2d + sha512h2 q0, q1, v2.2d + sha512su0 v11.2d, v12.2d + sha512su1 v11.2d, v13.2d, v14.2d + eor3 v25.16b, v12.16b, v7.16b, v2.16b + rax1 v30.2d, v29.2d, v26.2d + xar v26.2d, v21.2d, v27.2d, #63 + bcax v31.16b, v26.16b, v2.16b, v1.16b + +//CHECK-SHA: sha512h q0, q1, v2.2d // encoding: [0x20,0x80,0x62,0xce] +//CHECK-SHA: sha512h2 q0, q1, v2.2d // encoding: [0x20,0x84,0x62,0xce] +//CHECK-SHA: sha512su0 v11.2d, v12.2d // encoding: [0x8b,0x81,0xc0,0xce] +//CHECK-SHA: sha512su1 v11.2d, v13.2d, v14.2d // encoding: [0xab,0x89,0x6e,0xce] +//CHECK-SHA: eor3 v25.16b, v12.16b, v7.16b, v2.16b // encoding: [0x99,0x09,0x07,0xce] +//CHECK-SHA: rax1 v30.2d, v29.2d, v26.2d // encoding: [0xbe,0x8f,0x7a,0xce] +//CHECK-SHA: xar v26.2d, v21.2d, v27.2d, #63 // encoding: [0xba,0xfe,0x9b,0xce] +//CHECK-SHA: bcax v31.16b, v26.16b, v2.16b, v1.16b // encoding: [0x5f,0x07,0x22,0xce] + + +// CHECK-NO-SHA: error: instruction requires: sha3 +// CHECK-NO-SHA: error: instruction requires: sha3 +// CHECK-NO-SHA: error: instruction requires: sha3 +// CHECK-NO-SHA: error: instruction requires: sha3 +// CHECK-NO-SHA: error: instruction requires: sha3 +// CHECK-NO-SHA: error: instruction requires: sha3 +// CHECK-NO-SHA: error: instruction requires: sha3 +// CHECK-NO-SHA: error: instruction requires: sha3 + + sm3ss1 v20.4s, v23.4s, v21.4s, v22.4s + sm3tt1a v20.4s, v23.4s, v21.s[3] + sm3tt1b v20.4s, v23.4s, v21.s[3] + sm3tt2a v20.4s, v23.4s, v21.s[3] + sm3tt2b v20.4s, v23.4s, v21.s[3] + sm3partw1 v30.4s, v29.4s, v26.4s + sm3partw2 v30.4s, v29.4s, v26.4s + sm4ekey v11.4s, v11.4s, v19.4s + sm4e v2.4s, v15.4s + +// CHECK-SM: sm3ss1 v20.4s, v23.4s, v21.4s, v22.4s // encoding: [0xf4,0x5a,0x55,0xce] +// CHECK-SM: sm3tt1a v20.4s, v23.4s, v21.s[3] // encoding: [0xf4,0xb2,0x55,0xce] +// CHECK-SM: sm3tt1b v20.4s, v23.4s, v21.s[3] // encoding: [0xf4,0xb6,0x55,0xce] +// CHECK-SM: sm3tt2a v20.4s, v23.4s, v21.s[3] // encoding: [0xf4,0xba,0x55,0xce] +// CHECK-SM: sm3tt2b v20.4s, v23.4s, v21.s[3] // encoding: [0xf4,0xbe,0x55,0xce] +// CHECK-SM: sm3partw1 v30.4s, v29.4s, v26.4s // encoding: [0xbe,0xc3,0x7a,0xce] +// CHECK-SM: sm3partw2 v30.4s, v29.4s, v26.4s // encoding: [0xbe,0xc7,0x7a,0xce] +// CHECK-SM: sm4ekey v11.4s, v11.4s, v19.4s // encoding: [0x6b,0xc9,0x73,0xce] +// CHECK-SM: sm4e v2.4s, v15.4s // encoding: [0xe2,0x85,0xc0,0xce] + +// CHECK-NO-SM: error: instruction requires: sm4 +// CHECK-NO-SM: error: instruction requires: sm4 +// CHECK-NO-SM: error: instruction requires: sm4 +// CHECK-NO-SM: error: instruction requires: sm4 +// CHECK-NO-SM: error: instruction requires: sm4 +// CHECK-NO-SM: error: instruction requires: sm4 +// CHECK-NO-SM: error: instruction requires: sm4 diff --git a/llvm/test/MC/AArch64/directive-cpu-err.s b/llvm/test/MC/AArch64/directive-cpu-err.s index 134251e..235fbca 100644 --- a/llvm/test/MC/AArch64/directive-cpu-err.s +++ b/llvm/test/MC/AArch64/directive-cpu-err.s @@ -28,7 +28,7 @@ .cpu generic+nocrypto+crc aesd v0.16b, v2.16b - // CHECK: error: instruction requires: crypto + // CHECK: error: instruction requires: aes // CHECK-NEXT: aesd v0.16b, v2.16b // CHECK-NEXT: ^ @@ -42,3 +42,27 @@ // CHECK: error: unsupported architectural extension // CHECK-NEXT: .cpu generic+v8.1-a // CHECK-NEXT: ^ + + .cpu generic+noaes + aese v0.16b, v1.16b + // CHECK: error: instruction requires: aes + // CHECK-NEXT: aese v0.16b, v1.16b + // CHECK-NEXT: ^ + + .cpu generic+nosha2 + sha1h s0, s1 + // CHECK: error: instruction requires: sha2 + // CHECK-NEXT: sha1h s0, s1 + // CHECK-NEXT: ^ + + .cpu generic+nosha3 + sha512h q0, q1, v2.2d + // CHECK: error: instruction requires: sha3 + // CHECK-NEXT: sha512h q0, q1, v2.2d + // CHECK-NEXT: ^ + + .cpu generic+nosm4 + sm4e v2.4s, v15.4s + // CHECK: error: instruction requires: sm4 + // CHECK-NEXT: sm4e v2.4s, v15.4s + // CHECK-NEXT: ^ diff --git a/llvm/test/MC/AArch64/directive-cpu.s b/llvm/test/MC/AArch64/directive-cpu.s index 52efe06..9dc5631 100644 --- a/llvm/test/MC/AArch64/directive-cpu.s +++ b/llvm/test/MC/AArch64/directive-cpu.s @@ -21,11 +21,27 @@ aesd v0.16b, v2.16b .cpu generic+lse - casa w5, w7, [x20] + casa w5, w7, [x20] // CHECK: fminnm d0, d0, d1 // CHECK: fminnm d0, d0, d1 // CHECK: addp v0.4s, v0.4s, v0.4s // CHECK: crc32cx w0, w1, x3 // CHECK: aesd v0.16b, v2.16b -// CHECK: casa w5, w7, [x20] +// CHECK: casa w5, w7, [x20] + + .cpu generic+aes + aese v0.16b, v1.16b +// CHECK: aese v0.16b, v1.16b + + .cpu generic+sha2 + sha1h s0, s1 +// CHECK: sha1h s0, s1 + + .cpu generic+sha3 + sha512h q0, q1, v2.2d +// CHECK: sha512h q0, q1, v2.2d + + .cpu generic+sm4 + sm4e v2.4s, v15.4s +// CHECK: sm4e v2.4s, v15.4s diff --git a/llvm/test/MC/AArch64/neon-crypto.s b/llvm/test/MC/AArch64/neon-crypto.s index ed1bf88..0c53909 100644 --- a/llvm/test/MC/AArch64/neon-crypto.s +++ b/llvm/test/MC/AArch64/neon-crypto.s @@ -13,19 +13,19 @@ aesimc v0.16b, v1.16b // CHECK-NO-CRYPTO: error: instruction requires a CPU feature not currently enabled -// CHECK-NO-CRYPTO-ARM64: error: instruction requires: crypto +// CHECK-NO-CRYPTO-ARM64: error: instruction requires: aes // CHECK: aese v0.16b, v1.16b // encoding: [0x20,0x48,0x28,0x4e] // CHECK: aesd v0.16b, v1.16b // encoding: [0x20,0x58,0x28,0x4e] // CHECK: aesmc v0.16b, v1.16b // encoding: [0x20,0x68,0x28,0x4e] -// CHECK: aesimc v0.16b, v1.16b // encoding: [0x20,0x78,0x28,0x4e] +// CHECK: aesimc v0.16b, v1.16b // encoding: [0x20,0x78,0x28,0x4e] sha1h s0, s1 sha1su1 v0.4s, v1.4s sha256su0 v0.4s, v1.4s // CHECK: sha1h s0, s1 // encoding: [0x20,0x08,0x28,0x5e] -// CHECK: sha1su1 v0.4s, v1.4s // encoding: [0x20,0x18,0x28,0x5e] -// CHECK: sha256su0 v0.4s, v1.4s // encoding: [0x20,0x28,0x28,0x5e] +// CHECK: sha1su1 v0.4s, v1.4s // encoding: [0x20,0x18,0x28,0x5e] +// CHECK: sha256su0 v0.4s, v1.4s // encoding: [0x20,0x28,0x28,0x5e] sha1c q0, s1, v2.4s sha1p q0, s1, v2.4s @@ -38,8 +38,7 @@ // CHECK: sha1c q0, s1, v2.4s // encoding: [0x20,0x00,0x02,0x5e] // CHECK: sha1p q0, s1, v2.4s // encoding: [0x20,0x10,0x02,0x5e] // CHECK: sha1m q0, s1, v2.4s // encoding: [0x20,0x20,0x02,0x5e] -// CHECK: sha1su0 v0.4s, v1.4s, v2.4s // encoding: [0x20,0x30,0x02,0x5e] -// CHECK: sha256h q0, q1, v2.4s // encoding: [0x20,0x40,0x02,0x5e] -// CHECK: sha256h2 q0, q1, v2.4s // encoding: [0x20,0x50,0x02,0x5e] +// CHECK: sha1su0 v0.4s, v1.4s, v2.4s // encoding: [0x20,0x30,0x02,0x5e] +// CHECK: sha256h q0, q1, v2.4s // encoding: [0x20,0x40,0x02,0x5e] +// CHECK: sha256h2 q0, q1, v2.4s // encoding: [0x20,0x50,0x02,0x5e] // CHECK: sha256su1 v0.4s, v1.4s, v2.4s // encoding: [0x20,0x60,0x02,0x5e] - diff --git a/llvm/test/MC/AArch64/nofp-crypto-diagnostic.s b/llvm/test/MC/AArch64/nofp-crypto-diagnostic.s index 36da8a8..53e689a 100644 --- a/llvm/test/MC/AArch64/nofp-crypto-diagnostic.s +++ b/llvm/test/MC/AArch64/nofp-crypto-diagnostic.s @@ -3,6 +3,6 @@ sha1h s0, s1 -// CHECK-ERROR: error: instruction requires: crypto +// CHECK-ERROR: error: instruction requires: sha2 // CHECK-ERROR-NEXT: sha1h s0, s1 // CHECK-ERROR-NEXT: ^ diff --git a/llvm/test/MC/Disassembler/AArch64/armv8.2a-crypto.txt b/llvm/test/MC/Disassembler/AArch64/armv8.2a-crypto.txt new file mode 100644 index 0000000..bfa8797 --- /dev/null +++ b/llvm/test/MC/Disassembler/AArch64/armv8.2a-crypto.txt @@ -0,0 +1,93 @@ +# RUN: llvm-mc -triple aarch64-none-linux-gnu -mattr=+v8.4a,+crypto,+sm4,+sha3 --disassemble < %s | FileCheck %s +# RUN: llvm-mc -triple aarch64-none-linux-gnu -mattr=+v8.4a,-crypto,-sm4,-sha3 --disassemble < %s 2>&1 | FileCheck %s --check-prefix=CHECK-ERROR + +0x20,0x80,0x62,0xce +0x20,0x84,0x62,0xce +0x8b,0x81,0xc0,0xce +0xab,0x89,0x6e,0xce + +# CHECK: sha512h q0, q1, v2.2d +# CHECK: sha512h2 q0, q1, v2.2d +# CHECK: sha512su0 v11.2d, v12.2d +# CHECK: sha512su1 v11.2d, v13.2d, v14.2d + +0x99,0x09,0x07,0xce +0xbe,0x8f,0x7a,0xce +0xba,0xfe,0x9b,0xce +0x5f,0x07,0x22,0xce +0xf4,0x5a,0x55,0xce +0xf4,0xb2,0x55,0xce +0xf4,0xb6,0x55,0xce +0xf4,0xba,0x55,0xce +0xf4,0xbe,0x55,0xce +0xbe,0xc3,0x7a,0xce +0xbe,0xc7,0x7a,0xce +0x6b,0xc9,0x73,0xce +0xe2,0x85,0xc0,0xce + +# CHECK: eor3 v25.16b, v12.16b, v7.16b, v2.16b +# CHECK: rax1 v30.2d, v29.2d, v26.2d +# CHECK: xar v26.2d, v21.2d, v27.2d, #63 +# CHECK: bcax v31.16b, v26.16b, v2.16b, v1.16b +# CHECK: sm3ss1 v20.4s, v23.4s, v21.4s, v22.4s +# CHECK: sm3tt1a v20.4s, v23.4s, v21.s[3] +# CHECK: sm3tt1b v20.4s, v23.4s, v21.s[3] +# CHECK: sm3tt2a v20.4s, v23.4s, v21.s[3] +# CHECK: sm3tt2b v20.4s, v23.4s, v21.s[3] +# CHECK: sm3partw1 v30.4s, v29.4s, v26.4s +# CHECK: sm3partw2 v30.4s, v29.4s, v26.4s +# CHECK: sm4ekey v11.4s, v11.4s, v19.4s +# CHECK: sm4e v2.4s, v15.4s + +#CHECK-ERROR: warning: invalid instruction encoding +#CHECK-ERROR: 0x20,0x80,0x62,0xce +#CHECK-ERROR: ^ +#CHECK-ERROR: warning: invalid instruction encoding +#CHECK-ERROR: 0x20,0x84,0x62,0xce +#CHECK-ERROR: ^ +#CHECK-ERROR: warning: invalid instruction encoding +#CHECK-ERROR: 0x8b,0x81,0xc0,0xce +#CHECK-ERROR: ^ +#CHECK-ERROR: warning: invalid instruction encoding +#CHECK-ERROR: 0xab,0x89,0x6e,0xce +#CHECK-ERROR: ^ +#CHECK-ERROR: warning: invalid instruction encoding +#CHECK-ERROR: 0x99,0x09,0x07,0xce +#CHECK-ERROR: ^ +#CHECK-ERROR: warning: invalid instruction encoding +#CHECK-ERROR: 0xbe,0x8f,0x7a,0xce +#CHECK-ERROR: ^ +#CHECK-ERROR: warning: invalid instruction encoding +#CHECK-ERROR: 0xba,0xfe,0x9b,0xce +#CHECK-ERROR: ^ +#CHECK-ERROR: warning: invalid instruction encoding +#CHECK-ERROR: 0x5f,0x07,0x22,0xce +#CHECK-ERROR: ^ +#CHECK-ERROR: warning: invalid instruction encoding +#CHECK-ERROR: 0xf4,0x5a,0x55,0xce +#CHECK-ERROR: ^ +#CHECK-ERROR: warning: invalid instruction encoding +#CHECK-ERROR: 0xf4,0xb2,0x55,0xce +#CHECK-ERROR: ^ +#CHECK-ERROR: warning: invalid instruction encoding +#CHECK-ERROR: 0xf4,0xb6,0x55,0xce +#CHECK-ERROR: ^ +#CHECK-ERROR: warning: invalid instruction encoding +#CHECK-ERROR: 0xf4,0xba,0x55,0xce +#CHECK-ERROR: ^ +#CHECK-ERROR: warning: invalid instruction encoding +#CHECK-ERROR: 0xf4,0xbe,0x55,0xce +#CHECK-ERROR: ^ +#CHECK-ERROR: warning: invalid instruction encoding +#CHECK-ERROR: 0xbe,0xc3,0x7a,0xce +#CHECK-ERROR: ^ +#CHECK-ERROR: warning: invalid instruction encoding +#CHECK-ERROR: 0xbe,0xc7,0x7a,0xce +#CHECK-ERROR: ^ +#CHECK-ERROR: warning: invalid instruction encoding +#CHECK-ERROR: 0x6b,0xc9,0x73,0xce +#CHECK-ERROR: ^ +#CHECK-ERROR: warning: invalid instruction encoding +#CHECK-ERROR: 0xe2,0x85,0xc0,0xce +#CHECK-ERROR: ^ + -- 2.7.4