arch-arm: AArch32 Crypto SHA

This patch implements the AArch32 secure hashing instructions
from the Crypto extension.

Change-Id: Iaba8424ab71800228a9aff039d93f5c35ee7d8e5
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/13247
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>

9 files changed, 692 insertions, 1 deletions

diff --git a/src/arch/arm/SConscript b/src/arch/arm/SConscript
index d9b5661ca..c84c69ba6 100644
--- a/src/arch/arm/SConscript
+++ b/src/arch/arm/SConscript
@@ -1,6 +1,6 @@
 # -*- mode:python -*-
 
-# Copyright (c) 2009, 2012-2013 ARM Limited
+# Copyright (c) 2009, 2012-2013, 2018 ARM Limited
 # All rights reserved.
 #
 # The license below extends only to copyright in the software and shall
@@ -62,6 +62,7 @@ if env['TARGET_ISA'] == 'arm':
     Source('insts/static_inst.cc')
     Source('insts/vfp.cc')
     Source('insts/fplib.cc')
+    Source('insts/crypto.cc')
     Source('interrupts.cc')
     Source('isa.cc')
     Source('isa_device.cc')
diff --git a/src/arch/arm/insts/crypto.cc b/src/arch/arm/insts/crypto.cc
new file mode 100644
index 000000000..e5a4612bc
--- /dev/null
+++ b/src/arch/arm/insts/crypto.cc
@@ -0,0 +1,290 @@
+/*
+ * Copyright (c) 2018 ARM Limited
+ * All rights reserved
+ *
+ * The license below extends only to copyright in the software and shall
+ * not be construed as granting a license to any other intellectual
+ * property including but not limited to intellectual property relating
+ * to a hardware implementation of the functionality of the software
+ * licensed hereunder.  You may use the software subject to the license
+ * terms below provided that you ensure that this notice is replicated
+ * unmodified and in its entirety in all distributions of the software,
+ * modified or unmodified, in source code or in binary form.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are
+ * met: redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer;
+ * redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution;
+ * neither the name of the copyright holders nor the names of its
+ * contributors may be used to endorse or promote products derived from
+ * this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * Authors: Matt Horsnell
+ *          Prakash Ramrakhyani
+ */
+
+#include <cstdio>
+#include <iostream>
+#include <string>
+
+#include "crypto.hh"
+
+namespace ArmISA {
+
+void
+Crypto::sha256Op(
+    uint32_t *X,
+    uint32_t *Y,
+    uint32_t *Z)
+{
+    uint32_t T0, T1, T2, T3;
+    for (int i = 0; i < 4; ++i) {
+        T0 = choose(Y[0], Y[1], Y[2]);
+        T1 = majority(X[0], X[1], X[2]);
+        T2 = Y[3] + sigma1(Y[0]) + T0 + Z[i];
+        X[3] = T2 + X[3];
+        Y[3] = T2 + sigma0(X[0]) + T1;
+        // Rotate
+        T3 = Y[3];
+        Y[3] = Y[2]; Y[2] = Y[1]; Y[1] = Y[0]; Y[0] = X[3];
+        X[3] = X[2]; X[2] = X[1]; X[1] = X[0]; X[0] = T3;
+    }
+}
+
+void
+Crypto::_sha1Op(
+    uint32_t *X,
+    uint32_t *Y,
+    uint32_t *Z,
+    SHAOp op)
+{
+    uint32_t T1, T2;
+
+    for (int i = 0; i < 4; ++i) {
+        switch (op) {
+          case CHOOSE:   T1 = choose(X[1], X[2], X[3]); break;
+          case PARITY:   T1 = parity(X[1], X[2], X[3]); break;
+          case MAJORITY: T1 = majority(X[1], X[2], X[3]); break;
+          default: return;
+        }
+        Y[0] += ror(X[0], 27) + T1 + Z[i];
+        X[1] = ror(X[1], 2);
+        T2 = Y[0];
+        Y[0] = X[3];
+        X[3] = X[2]; X[2] = X[1]; X[1] = X[0]; X[0] = T2;
+    }
+}
+
+void
+Crypto::sha256H(
+    uint8_t *output,
+    uint8_t *input,
+    uint8_t *input2)
+{
+    uint32_t X[4], Y[4], Z[4];
+    load3Reg(&X[0], &Y[0], &Z[0], output, input, input2);
+    sha256Op(&X[0], &Y[0], &Z[0]);
+    store1Reg(output, &X[0]);
+}
+
+void
+Crypto::sha256H2(
+    uint8_t *output,
+    uint8_t *input,
+    uint8_t *input2)
+{
+    uint32_t X[4], Y[4], Z[4];
+    load3Reg(&X[0], &Y[0], &Z[0], output, input, input2);
+    sha256Op(&Y[0], &X[0], &Z[0]);
+    store1Reg(output, &X[0]);
+}
+
+void
+Crypto::sha256Su0(uint8_t *output, uint8_t *input)
+{
+    uint32_t X[4], Y[4];
+    uint32_t T[4];
+
+    load2Reg(&X[0], &Y[0], output, input);
+
+    T[3] = Y[0]; T[2] = X[3]; T[1] = X[2]; T[0] = X[1];
+
+    T[3] = ror(T[3], 7) ^ ror(T[3], 18) ^ (T[3] >> 3);
+    T[2] = ror(T[2], 7) ^ ror(T[2], 18) ^ (T[2] >> 3);
+    T[1] = ror(T[1], 7) ^ ror(T[1], 18) ^ (T[1] >> 3);
+    T[0] = ror(T[0], 7) ^ ror(T[0], 18) ^ (T[0] >> 3);
+
+    X[3] += T[3];
+    X[2] += T[2];
+    X[1] += T[1];
+    X[0] += T[0];
+
+    store1Reg(output, &X[0]);
+}
+
+void
+Crypto::sha256Su1(
+    uint8_t *output,
+    uint8_t *input,
+    uint8_t *input2)
+{
+    uint32_t X[4], Y[4], Z[4];
+    uint32_t T0[4], T1[4], T2[4], T3[4];
+
+    load3Reg(&X[0], &Y[0], &Z[0], output, input, input2);
+
+    T0[3] = Z[0]; T0[2] = Y[3]; T0[1] = Y[2]; T0[0] = Y[1];
+    T1[1] = Z[3]; T1[0] = Z[2];
+    T1[1] = ror(T1[1], 17) ^ ror(T1[1], 19) ^ (T1[1] >> 10);
+    T1[0] = ror(T1[0], 17) ^ ror(T1[0], 19) ^ (T1[0] >> 10);
+    T3[1] = X[1] + T0[1]; T3[0] = X[0] + T0[0];
+    T1[1] = T3[1] + T1[1]; T1[0] = T3[0] + T1[0];
+    T2[1] = ror(T1[1], 17) ^ ror(T1[1], 19) ^ (T1[1] >> 10);
+    T2[0] = ror(T1[0], 17) ^ ror(T1[0], 19) ^ (T1[0] >> 10);
+    T3[1] = X[3] + T0[3]; T3[0] = X[2] + T0[2];
+    X[3] = T3[1] + T2[1];
+    X[2] = T3[0] + T2[0];
+    X[1] = T1[1]; X[0] = T1[0];
+
+    store1Reg(output, &X[0]);
+}
+
+void
+Crypto::sha1Op(
+    uint8_t *output,
+    uint8_t *input,
+    uint8_t *input2,
+    SHAOp op)
+{
+    uint32_t X[4], Y[4], Z[4];
+    load3Reg(&X[0], &Y[0], &Z[0], output, input, input2);
+    _sha1Op(&X[0], &Y[0], &Z[0], op);
+    store1Reg(output, &X[0]);
+}
+
+void
+Crypto::sha1C(
+    uint8_t *output,
+    uint8_t *input,
+    uint8_t *input2)
+{
+    sha1Op(output, input, input2, CHOOSE);
+}
+
+void
+Crypto::sha1P(
+    uint8_t *output,
+    uint8_t *input,
+    uint8_t *input2)
+{
+    sha1Op(output, input, input2, PARITY);
+}
+
+void
+Crypto::sha1M(
+    uint8_t *output,
+    uint8_t *input,
+    uint8_t *input2)
+{
+    sha1Op(output, input, input2, MAJORITY);
+}
+
+void
+Crypto::sha1H(uint8_t *output, uint8_t *input)
+{
+    uint32_t X[4], Y[4];
+    load2Reg(&X[0], &Y[0], output, input);
+    X[0] = ror(Y[0], 2);
+    store1Reg(output, &X[0]);
+}
+
+void
+Crypto::sha1Su0(
+    uint8_t *output,
+    uint8_t *input,
+    uint8_t *input2)
+{
+    uint32_t X[4], Y[4], Z[4], T[4];
+    load3Reg(&X[0], &Y[0], &Z[0], output, input, input2);
+
+    T[3] = Y[1]; T[2] = Y[0]; T[1] = X[3]; T[0] = X[2];
+    X[3] = T[3] ^ X[3] ^ Z[3];
+    X[2] = T[2] ^ X[2] ^ Z[2];
+    X[1] = T[1] ^ X[1] ^ Z[1];
+    X[0] = T[0] ^ X[0] ^ Z[0];
+
+    store1Reg(output, &X[0]);
+}
+
+void
+Crypto::sha1Su1(uint8_t *output, uint8_t *input)
+{
+    uint32_t X[4], Y[4], T[4];
+    load2Reg(&X[0], &Y[0], output, input);
+
+    T[3] = X[3] ^ 0x0;
+    T[2] = X[2] ^ Y[3];
+    T[1] = X[1] ^ Y[2];
+    T[0] = X[0] ^ Y[1];
+    X[2] = ror(T[2], 31); X[1] = ror(T[1], 31); X[0] = ror(T[0], 31);
+    X[3] = ror(T[3], 31) ^ ror(T[0], 30);
+
+    store1Reg(output, &X[0]);
+}
+
+void
+Crypto::load2Reg(
+    uint32_t *X,
+    uint32_t *Y,
+    uint8_t *output,
+    uint8_t *input)
+{
+    for (int i = 0; i < 4; ++i) {
+        X[i] = *((uint32_t *)&output[i*4]);
+        Y[i] = *((uint32_t *)&input[i*4]);
+    }
+}
+
+void
+Crypto::load3Reg(
+    uint32_t *X,
+    uint32_t *Y,
+    uint32_t *Z,
+    uint8_t *output,
+    uint8_t *input,
+    uint8_t *input2)
+{
+    for (int i = 0; i < 4; ++i) {
+        X[i] = *((uint32_t *)&output[i*4]);
+        Y[i] = *((uint32_t *)&input[i*4]);
+        Z[i] = *((uint32_t *)&input2[i*4]);
+    }
+}
+
+void
+Crypto::store1Reg(uint8_t *output, uint32_t *X)
+{
+    for (int i = 0; i < 4; ++i) {
+        output[i*4] = (uint8_t)(X[i]);
+        output[i*4+1] = (uint8_t)(X[i] >> 8);
+        output[i*4+2] = (uint8_t)(X[i] >> 16);
+        output[i*4+3] = (uint8_t)(X[i] >> 24);
+    }
+}
+
+} // namespace ArmISA
diff --git a/src/arch/arm/insts/crypto.hh b/src/arch/arm/insts/crypto.hh
new file mode 100644
index 000000000..c89ace04a
--- /dev/null
+++ b/src/arch/arm/insts/crypto.hh
@@ -0,0 +1,110 @@
+/*
+ * Copyright (c) 2018 ARM Limited
+ * All rights reserved
+ *
+ * The license below extends only to copyright in the software and shall
+ * not be construed as granting a license to any other intellectual
+ * property including but not limited to intellectual property relating
+ * to a hardware implementation of the functionality of the software
+ * licensed hereunder.  You may use the software subject to the license
+ * terms below provided that you ensure that this notice is replicated
+ * unmodified and in its entirety in all distributions of the software,
+ * modified or unmodified, in source code or in binary form.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are
+ * met: redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer;
+ * redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution;
+ * neither the name of the copyright holders nor the names of its
+ * contributors may be used to endorse or promote products derived from
+ * this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * Authors: Matt Horsnell
+ *          Prakash Ramrakhyani
+ */
+
+#ifndef __ARCH_ARM_INSTS_CRYPTO_HH__
+#define __ARCH_ARM_INSTS_CRYPTO_HH__
+
+namespace ArmISA {
+
+class Crypto
+{
+    enum SHAOp : uint8_t
+    {
+        CHOOSE = 0,
+        PARITY,
+        MAJORITY
+    };
+
+    uint32_t ror(uint32_t x, uint8_t shift)
+    {
+        return (x >> shift) | (x << (32 - shift));
+    }
+
+    uint32_t choose(uint32_t X, uint32_t Y, uint32_t Z)
+    {
+        return (((Y ^ Z) & X) ^ Z);
+    }
+
+    uint32_t parity(uint32_t X, uint32_t Y, uint32_t Z)
+    {
+        return (X ^ Y ^ Z);
+    }
+
+    uint32_t majority(uint32_t X, uint32_t Y, uint32_t Z)
+    {
+        return ((X & Y) | ((X | Y) & Z));
+    }
+
+    uint32_t sigma0(uint32_t X)
+    {
+        return ror(X,2) ^ ror(X,13) ^ ror(X,22);
+    }
+
+    uint32_t sigma1(uint32_t X)
+    {
+        return ror(X,6) ^ ror(X,11) ^ ror(X,25);
+    }
+
+    void sha256Op(uint32_t *X, uint32_t *Y, uint32_t *Z);
+    void sha1Op(uint8_t *output, uint8_t *input, uint8_t *input2, SHAOp op);
+    void _sha1Op(uint32_t *X, uint32_t *Y, uint32_t *Z, SHAOp op);
+
+    void load2Reg(uint32_t *X, uint32_t *Y, uint8_t *output, uint8_t *input);
+    void load3Reg(uint32_t *X, uint32_t *Y, uint32_t *Z,
+                  uint8_t *output, uint8_t *input, uint8_t *input2);
+    void store1Reg(uint8_t *output, uint32_t *X);
+
+  public:
+    void sha256H(uint8_t *output, uint8_t *input, uint8_t *input2);
+    void sha256H2(uint8_t *output, uint8_t *input, uint8_t *input2);
+    void sha256Su0(uint8_t *output, uint8_t *input);
+    void sha256Su1(uint8_t *output, uint8_t *input, uint8_t *input2);
+
+    void sha1C(uint8_t *output, uint8_t *input, uint8_t *input2);
+    void sha1P(uint8_t *output, uint8_t *input, uint8_t *input2);
+    void sha1M(uint8_t *output, uint8_t *input, uint8_t *input2);
+    void sha1H(uint8_t *output, uint8_t *input);
+    void sha1Su0(uint8_t *output, uint8_t *input, uint8_t *input2);
+    void sha1Su1(uint8_t *output, uint8_t *input);
+};
+
+} // namespace ArmISA
+
+#endif //__ARCH_ARM_INSTS_CRYPTO_HH__
diff --git a/src/arch/arm/isa/formats/fp.isa b/src/arch/arm/isa/formats/fp.isa
index 2412a1f10..26abc659f 100644
--- a/src/arch/arm/isa/formats/fp.isa
+++ b/src/arch/arm/isa/formats/fp.isa
@@ -606,6 +606,34 @@ let {{
                         }
                     }
                 }
+            } else {
+                if (u) {
+                    switch (c) {
+                      case 0x0:
+                        return new SHA256H(machInst, vd, vn, vm);
+                      case 0x1:
+                        return new SHA256H2(machInst, vd, vn, vm);
+                      case 0x2:
+                        return new SHA256SU1(machInst, vd, vn, vm);
+                      case 0x3:
+                        return new Unknown(machInst);
+                      default:
+                        M5_UNREACHABLE;
+                    }
+                } else {
+                    switch (c) {
+                      case 0x0:
+                        return new SHA1C(machInst, vd, vn, vm);
+                      case 0x1:
+                        return new SHA1P(machInst, vd, vn, vm);
+                      case 0x2:
+                        return new SHA1M(machInst, vd, vn, vm);
+                      case 0x3:
+                        return new SHA1SU0(machInst, vd, vn, vm);
+                      default:
+                        M5_UNREACHABLE;
+                    }
+                }
             }
             return new Unknown(machInst);
           case 0xd:
@@ -1534,6 +1562,12 @@ let {{
                     return decodeNeonSTwoMiscReg<NVcltD, NVcltQ>(
                             q, size, machInst, vd, vm);
                 }
+              case 0x5:
+                if (q) {
+                    return new SHA1H(machInst, vd, vm);
+                } else {
+                    return new Unknown(machInst);
+                }
               case 0x6:
                 if (bits(machInst, 10)) {
                     if (q)
@@ -1596,6 +1630,12 @@ let {{
                 } else {
                     return new Unknown(machInst);
                 }
+              case 0x7:
+                if (q) {
+                    return new SHA256SU0(machInst, vd, vm);
+                } else {
+                    return new SHA1SU1(machInst, vd, vm);
+                }
               case 0xc:
               case 0xe:
                 if (b == 0x18) {
diff --git a/src/arch/arm/isa/includes.isa b/src/arch/arm/isa/includes.isa
index caba46de9..37578f6ce 100644
--- a/src/arch/arm/isa/includes.isa
+++ b/src/arch/arm/isa/includes.isa
@@ -51,6 +51,7 @@ output header {{
 
 #include "arch/arm/insts/branch.hh"
 #include "arch/arm/insts/branch64.hh"
+#include "arch/arm/insts/crypto.hh"
 #include "arch/arm/insts/data64.hh"
 #include "arch/arm/insts/fplib.hh"
 #include "arch/arm/insts/macromem.hh"
diff --git a/src/arch/arm/isa/insts/crypto.isa b/src/arch/arm/isa/insts/crypto.isa
new file mode 100644
index 000000000..cdc0293a4
--- /dev/null
+++ b/src/arch/arm/isa/insts/crypto.isa
@@ -0,0 +1,174 @@
+// -*- mode:c++ -*-
+//
+// Copyright (c) 2018 ARM Limited
+// All rights reserved
+//
+// The license below extends only to copyright in the software and shall
+// not be construed as granting a license to any other intellectual
+// property including but not limited to intellectual property relating
+// to a hardware implementation of the functionality of the software
+// licensed hereunder.  You may use the software subject to the license
+// terms below provided that you ensure that this notice is replicated
+// unmodified and in its entirety in all distributions of the software,
+// modified or unmodified, in source code or in binary form.
+//
+// Redistribution and use in source and binary forms, with or without
+// modification, are permitted provided that the following conditions are
+// met: redistributions of source code must retain the above copyright
+// notice, this list of conditions and the following disclaimer;
+// redistributions in binary form must reproduce the above copyright
+// notice, this list of conditions and the following disclaimer in the
+// documentation and/or other materials provided with the distribution;
+// neither the name of the copyright holders nor the names of its
+// contributors may be used to endorse or promote products derived from
+// this software without specific prior written permission.
+//
+// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+//
+// Authors: Matt Horsnell
+//          Prakash Ramrakhyani
+
+let {{
+
+    cryptoEnabledCheckCode = '''
+        auto crypto_reg = xc->tcBase()->readMiscReg(MISCREG_ID_ISAR5);
+        if (!(crypto_reg & %(mask)d)) {
+            return std::make_shared<UndefinedInstruction>(machInst, true);
+        }
+    '''
+
+    header_output = ""
+    decoder_output = ""
+    exec_output = ""
+
+    cryptoRegRegRegPrefix = '''
+        Crypto crypto;
+        RegVect srcReg1, srcReg2, destReg;
+        // Read source and destination registers.
+    '''
+    for reg in range(4):
+        cryptoRegRegRegPrefix += '''
+            srcReg1.regs[%(reg)d] = htog(FpOp1P%(reg)d_uw);
+            srcReg2.regs[%(reg)d] = htog(FpOp2P%(reg)d_uw);
+            destReg.regs[%(reg)d] = htog(FpDestP%(reg)d_uw);
+        ''' % { "reg" : reg }
+    cryptoRegRegRegPrefix += '''
+        unsigned char *output = (unsigned char *)(&destReg.regs[0]);
+        unsigned char *input  = (unsigned char *)(&srcReg1.regs[0]);
+        unsigned char *input2 = (unsigned char *)(&srcReg2.regs[0]);
+    '''
+
+    cryptoSuffix = ""
+    for reg in range(4):
+        cryptoSuffix += '''
+            FpDestP%(reg)d_uw = gtoh(destReg.regs[%(reg)d]);
+        ''' % { "reg" : reg }
+
+    cryptoRegRegPrefix = '''
+        Crypto crypto;
+        RegVect srcReg1, destReg;
+        // Read source and destination registers.
+    '''
+    for reg in range(4):
+        cryptoRegRegPrefix += '''
+            srcReg1.regs[%(reg)d] = htog(FpOp1P%(reg)d_uw);
+            destReg.regs[%(reg)d] = htog(FpDestP%(reg)d_uw);
+        ''' % { "reg" : reg }
+
+    cryptoRegRegPrefix += '''
+        // cast into format passed to aes encrypt method.
+        unsigned char *output = (unsigned char *)(&destReg.regs[0]);
+        unsigned char *input  = (unsigned char *)(&srcReg1.regs[0]);
+    '''
+
+    def cryptoRegRegRegInst(name, Name, opClass, enable_check, crypto_func):
+        global header_output, decoder_output, exec_output
+
+        crypto_prefix = enable_check + cryptoRegRegRegPrefix
+        cryptocode = crypto_prefix + crypto_func + cryptoSuffix
+
+        cryptoiop = InstObjParams(name, Name, "RegRegRegOp",
+                                { "code": cryptocode,
+                                  "r_count": 4,
+                                  "predicate_test": predicateTest,
+                                  "op_class": opClass}, [])
+        header_output += RegRegRegOpDeclare.subst(cryptoiop)
+        decoder_output += RegRegRegOpConstructor.subst(cryptoiop)
+        exec_output += CryptoPredOpExecute.subst(cryptoiop)
+
+    def cryptoRegRegInst(name, Name, opClass, enable_check, crypto_func):
+        global header_output, decoder_output, exec_output
+
+        crypto_prefix = enable_check + cryptoRegRegPrefix
+        cryptocode = crypto_prefix + crypto_func + cryptoSuffix
+
+        cryptoiop = InstObjParams(name, Name, "RegRegOp",
+                                { "code": cryptocode,
+                                  "r_count": 4,
+                                  "predicate_test": predicateTest,
+                                  "op_class": opClass}, [])
+        header_output += RegRegOpDeclare.subst(cryptoiop)
+        decoder_output += RegRegOpConstructor.subst(cryptoiop)
+        exec_output += CryptoPredOpExecute.subst(cryptoiop)
+
+    def cryptoRegRegImmInst(name, Name, opClass, enable_check, crypto_func):
+        global header_output, decoder_output, exec_output
+
+        crypto_prefix = enable_check + cryptoRegRegPrefix
+        cryptocode = crypto_prefix + crypto_func + cryptoSuffix
+
+        cryptoiop = InstObjParams(name, Name, "RegRegImmOp",
+                                { "code": cryptocode,
+                                  "r_count": 4,
+                                  "predicate_test": predicateTest,
+                                  "op_class": opClass}, [])
+        header_output += RegRegImmOpDeclare.subst(cryptoiop)
+        decoder_output += RegRegImmOpConstructor.subst(cryptoiop)
+        exec_output += CryptoPredOpExecute.subst(cryptoiop)
+
+    sha1_cCode = "crypto.sha1C(output, input, input2);"
+    sha1_pCode = "crypto.sha1P(output, input, input2);"
+    sha1_mCode = "crypto.sha1M(output, input, input2);"
+    sha1_hCode = "crypto.sha1H(output, input);"
+    sha1_su0Code = "crypto.sha1Su0(output, input, input2);"
+    sha1_su1Code = "crypto.sha1Su1(output, input);"
+
+    sha256_hCode = "crypto.sha256H(output, input, input2);"
+    sha256_h2Code = "crypto.sha256H2(output, input, input2);"
+    sha256_su0Code = "crypto.sha256Su0(output, input);"
+    sha256_su1Code = "crypto.sha256Su1(output, input, input2);"
+
+    sha1_enabled = cryptoEnabledCheckCode % { "mask" : 0xF00 }
+    cryptoRegRegRegInst("sha1c", "SHA1C", "SimdSha1HashOp",
+                        sha1_enabled, sha1_cCode)
+    cryptoRegRegRegInst("sha1p", "SHA1P", "SimdSha1HashOp",
+                        sha1_enabled, sha1_pCode)
+    cryptoRegRegRegInst("sha1m", "SHA1M", "SimdSha1HashOp",
+                        sha1_enabled, sha1_mCode)
+    cryptoRegRegInst("sha1h", "SHA1H", "SimdSha1Hash2Op",
+                     sha1_enabled, sha1_hCode)
+    cryptoRegRegRegInst("sha1su0", "SHA1SU0", "SimdShaSigma3Op",
+                        sha1_enabled, sha1_su0Code)
+    cryptoRegRegInst("sha1su1", "SHA1SU1", "SimdShaSigma2Op",
+                     sha1_enabled, sha1_su1Code)
+
+    sha2_enabled = cryptoEnabledCheckCode % { "mask" : 0xF000 }
+    cryptoRegRegRegInst("sha256h", "SHA256H", "SimdSha256HashOp",
+                        sha2_enabled, sha256_hCode)
+    cryptoRegRegRegInst("sha256h2", "SHA256H2", "SimdSha256Hash2Op",
+                        sha2_enabled, sha256_h2Code)
+    cryptoRegRegInst("sha256su0", "SHA256SU0", "SimdShaSigma2Op",
+                     sha2_enabled, sha256_su0Code)
+    cryptoRegRegRegInst("sha256su1", "SHA256SU1", "SimdShaSigma3Op",
+                        sha2_enabled, sha256_su1Code)
+}};
diff --git a/src/arch/arm/isa/insts/insts.isa b/src/arch/arm/isa/insts/insts.isa
index 89c80cbaa..b95356b70 100644
--- a/src/arch/arm/isa/insts/insts.isa
+++ b/src/arch/arm/isa/insts/insts.isa
@@ -102,3 +102,6 @@ split decoder;
 
 //m5 Pseudo-ops
 ##include "m5ops.isa"
+
+//Crypto
+##include "crypto.isa"
diff --git a/src/arch/arm/isa/templates/crypto.isa b/src/arch/arm/isa/templates/crypto.isa
new file mode 100644
index 000000000..083ef1336
--- /dev/null
+++ b/src/arch/arm/isa/templates/crypto.isa
@@ -0,0 +1,69 @@
+// -*- mode:c++ -*-
+
+// Copyright (c) 2018 ARM Limited
+// All rights reserved
+//
+// The license below extends only to copyright in the software and shall
+// not be construed as granting a license to any other intellectual
+// property including but not limited to intellectual property relating
+// to a hardware implementation of the functionality of the software
+// licensed hereunder.  You may use the software subject to the license
+// terms below provided that you ensure that this notice is replicated
+// unmodified and in its entirety in all distributions of the software,
+// modified or unmodified, in source code or in binary form.
+//
+// Redistribution and use in source and binary forms, with or without
+// modification, are permitted provided that the following conditions are
+// met: redistributions of source code must retain the above copyright
+// notice, this list of conditions and the following disclaimer;
+// redistributions in binary form must reproduce the above copyright
+// notice, this list of conditions and the following disclaimer in the
+// documentation and/or other materials provided with the distribution;
+// neither the name of the copyright holders nor the names of its
+// contributors may be used to endorse or promote products derived from
+// this software without specific prior written permission.
+//
+// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+//
+// Authors: Matt Horsnell
+
+// These currently all work on quad words, so some element/register
+// storage/extraction here is fixed as constants.
+def template CryptoPredOpExecute {{
+    Fault %(class_name)s::execute(ExecContext *xc,
+                                  Trace::InstRecord *traceData) const
+    {
+        Fault fault = NoFault;
+        %(op_decl)s;
+        %(op_rd)s;
+
+        const unsigned rCount = %(r_count)d;
+
+        union RegVect {
+            FloatRegBits regs[rCount];
+        };
+
+        if (%(predicate_test)s)
+        {
+            %(code)s;
+            if (fault == NoFault)
+            {
+                %(op_wb)s;
+            }
+        } else {
+            xc->setPredicate(false);
+        }
+
+        return fault;
+    }
+}};
diff --git a/src/arch/arm/isa/templates/templates.isa b/src/arch/arm/isa/templates/templates.isa
index 6d525738d..14913b358 100644
--- a/src/arch/arm/isa/templates/templates.isa
+++ b/src/arch/arm/isa/templates/templates.isa
@@ -76,4 +76,7 @@
 //Templates for Neon instructions
 ##include "neon.isa"
 
+//Templates for Crypto instructions
+##include "crypto.isa"
+
 ##include "neon64.isa"