From: Julian Seward Date: Fri, 15 Jul 2016 10:31:34 +0000 (+0000) Subject: Implement arm64 instructions: AESE AESD AESMC AESIMC. n-i-bz. X-Git-Tag: svn/VALGRIND_3_12_0^2~42 X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=bbd4a79d230f60ededdefbf4a458e544c645f66b;p=thirdparty%2Fvalgrind.git Implement arm64 instructions: AESE AESD AESMC AESIMC. n-i-bz. git-svn-id: svn://svn.valgrind.org/vex/trunk@3224 --- diff --git a/VEX/priv/guest_arm64_defs.h b/VEX/priv/guest_arm64_defs.h index 0d8b83b497..3cdeb4ce16 100644 --- a/VEX/priv/guest_arm64_defs.h +++ b/VEX/priv/guest_arm64_defs.h @@ -115,6 +115,15 @@ ULong arm64g_calculate_condition ( /* ARM64Condcode << 4 | cc_op */ extern ULong arm64g_dirtyhelper_MRS_CNTVCT_EL0 ( void ); +extern void arm64g_dirtyhelper_AESE ( /*OUT*/V128* res, + ULong argHi, ULong argLo ); +extern void arm64g_dirtyhelper_AESD ( /*OUT*/V128* res, + ULong argHi, ULong argLo ); +extern void arm64g_dirtyhelper_AESMC ( /*OUT*/V128* res, + ULong argHi, ULong argLo ); +extern void arm64g_dirtyhelper_AESIMC ( /*OUT*/V128* res, + ULong argHi, ULong argLo ); + /*---------------------------------------------------------*/ /*--- Condition code stuff ---*/ diff --git a/VEX/priv/guest_arm64_helpers.c b/VEX/priv/guest_arm64_helpers.c index 0cf804608b..e102e02dc4 100644 --- a/VEX/priv/guest_arm64_helpers.c +++ b/VEX/priv/guest_arm64_helpers.c @@ -692,6 +692,230 @@ ULong arm64g_dirtyhelper_MRS_CNTVCT_EL0 ( void ) } +/* DIRTY HELPERS for doing AES support: + * AESE (SubBytes, then ShiftRows) + * AESD (InvShiftRows, then InvSubBytes) + * AESMC (MixColumns) + * AESIMC (InvMixColumns) + These don't actually have to be dirty helpers -- they could be + clean, but for the fact that they return a V128 and a clean helper + can't do that. + + The ARMv8 manual seems to imply that AESE first performs ShiftRows, + then SubBytes. This seems to contradict FIPS 197, so the + implementation below is consistent with FIPS 197. One can observe + that the two transformations commute -- the order in which they + happen makes no difference to the result. So the ambiguity doesn't + actually matter, but it is confusing. The v8 manual looks correct + about AESD, though. + + The three functions rj_xtime, aesMixColumn and aesInvMixColumn only, + are taken from "A byte-oriented AES-256 implementation" and are subject + to the following usage terms: + + Byte-oriented AES-256 implementation. + All lookup tables replaced with 'on the fly' calculations. + + Copyright (c) 2007-2011 Ilya O. Levin, http://www.literatecode.com + Other contributors: Hal Finney + + Permission to use, copy, modify, and distribute this software for any + purpose with or without fee is hereby granted, provided that the above + copyright notice and this permission notice appear in all copies. + + THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES + WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF + MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR + ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES + WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN + ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF + OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. +*/ + +const UChar aesMapSubBytes[256] + = { 0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5, + 0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76, + 0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0, + 0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0, + 0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc, + 0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15, + 0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a, + 0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75, + 0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0, + 0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84, + 0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b, + 0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf, + 0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85, + 0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8, + 0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5, + 0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2, + 0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17, + 0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73, + 0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88, + 0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb, + 0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c, + 0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79, + 0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9, + 0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08, + 0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6, + 0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a, + 0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e, + 0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e, + 0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94, + 0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf, + 0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68, + 0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16 + }; + +const UChar aesMapInvSubBytes[256] + = { 0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38, + 0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb, + 0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87, + 0x34, 0x8e, 0x43, 0x44, 0xc4, 0xde, 0xe9, 0xcb, + 0x54, 0x7b, 0x94, 0x32, 0xa6, 0xc2, 0x23, 0x3d, + 0xee, 0x4c, 0x95, 0x0b, 0x42, 0xfa, 0xc3, 0x4e, + 0x08, 0x2e, 0xa1, 0x66, 0x28, 0xd9, 0x24, 0xb2, + 0x76, 0x5b, 0xa2, 0x49, 0x6d, 0x8b, 0xd1, 0x25, + 0x72, 0xf8, 0xf6, 0x64, 0x86, 0x68, 0x98, 0x16, + 0xd4, 0xa4, 0x5c, 0xcc, 0x5d, 0x65, 0xb6, 0x92, + 0x6c, 0x70, 0x48, 0x50, 0xfd, 0xed, 0xb9, 0xda, + 0x5e, 0x15, 0x46, 0x57, 0xa7, 0x8d, 0x9d, 0x84, + 0x90, 0xd8, 0xab, 0x00, 0x8c, 0xbc, 0xd3, 0x0a, + 0xf7, 0xe4, 0x58, 0x05, 0xb8, 0xb3, 0x45, 0x06, + 0xd0, 0x2c, 0x1e, 0x8f, 0xca, 0x3f, 0x0f, 0x02, + 0xc1, 0xaf, 0xbd, 0x03, 0x01, 0x13, 0x8a, 0x6b, + 0x3a, 0x91, 0x11, 0x41, 0x4f, 0x67, 0xdc, 0xea, + 0x97, 0xf2, 0xcf, 0xce, 0xf0, 0xb4, 0xe6, 0x73, + 0x96, 0xac, 0x74, 0x22, 0xe7, 0xad, 0x35, 0x85, + 0xe2, 0xf9, 0x37, 0xe8, 0x1c, 0x75, 0xdf, 0x6e, + 0x47, 0xf1, 0x1a, 0x71, 0x1d, 0x29, 0xc5, 0x89, + 0x6f, 0xb7, 0x62, 0x0e, 0xaa, 0x18, 0xbe, 0x1b, + 0xfc, 0x56, 0x3e, 0x4b, 0xc6, 0xd2, 0x79, 0x20, + 0x9a, 0xdb, 0xc0, 0xfe, 0x78, 0xcd, 0x5a, 0xf4, + 0x1f, 0xdd, 0xa8, 0x33, 0x88, 0x07, 0xc7, 0x31, + 0xb1, 0x12, 0x10, 0x59, 0x27, 0x80, 0xec, 0x5f, + 0x60, 0x51, 0x7f, 0xa9, 0x19, 0xb5, 0x4a, 0x0d, + 0x2d, 0xe5, 0x7a, 0x9f, 0x93, 0xc9, 0x9c, 0xef, + 0xa0, 0xe0, 0x3b, 0x4d, 0xae, 0x2a, 0xf5, 0xb0, + 0xc8, 0xeb, 0xbb, 0x3c, 0x83, 0x53, 0x99, 0x61, + 0x17, 0x2b, 0x04, 0x7e, 0xba, 0x77, 0xd6, 0x26, + 0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d + }; + +static inline UChar rj_xtime ( UChar x ) +{ + UChar y = (UChar)(x << 1); + return (x & 0x80) ? (y ^ 0x1b) : y; +} + +static void aesMixColumn ( /*MOD*/UChar* r ) +{ + UChar a = r[0]; + UChar b = r[1]; + UChar c = r[2]; + UChar d = r[3]; + UChar e = a ^ b ^ c ^ d; + r[0] ^= e ^ rj_xtime(a ^ b); + r[1] ^= e ^ rj_xtime(b ^ c); + r[2] ^= e ^ rj_xtime(c ^ d); + r[3] ^= e ^ rj_xtime(d ^ a); +} + +static void aesInvMixColumn ( /*MOD*/UChar* r ) +{ + UChar a = r[0]; + UChar b = r[1]; + UChar c = r[2]; + UChar d = r[3]; + UChar e = a ^ b ^ c ^ d; + UChar z = rj_xtime(e); + UChar x = e ^ rj_xtime(rj_xtime(z ^ a ^ c)); + UChar y = e ^ rj_xtime(rj_xtime(z ^ b ^ d)); + r[0] ^= x ^ rj_xtime(a ^ b); + r[1] ^= y ^ rj_xtime(b ^ c); + r[2] ^= x ^ rj_xtime(c ^ d); + r[3] ^= y ^ rj_xtime(d ^ a); +} + + +/* CALLED FROM GENERATED CODE */ +void arm64g_dirtyhelper_AESE ( /*OUT*/V128* res, ULong argHi, ULong argLo ) +{ + res->w64[1] = argHi; + res->w64[0] = argLo; + + /* First do SubBytes on the State. */ + UInt i; + for (i = 0; i < 16; i++) { + res->w8[i] = aesMapSubBytes[res->w8[i] & 0xFF]; + } + + /* Then do ShiftRows on the State. */ +# define XX(_ix) res->w8[_ix] + { UChar old1 = XX(1); + XX(1) = XX(5); XX(5) = XX(9); XX(9) = XX(13); XX(13) = old1; + } + { UChar old2 = XX(2); UChar old6 = XX(6); + XX(2) = XX(10); XX(6) = XX(14); XX(10) = old2; XX(14) = old6; + } + { UChar old15 = XX(15); + XX(15) = XX(11); XX(11) = XX(7); XX(7) = XX(3); XX(3) = old15; + } +# undef XX +} + + +/* CALLED FROM GENERATED CODE */ +void arm64g_dirtyhelper_AESD ( /*OUT*/V128* res, ULong argHi, ULong argLo ) +{ + res->w64[1] = argHi; + res->w64[0] = argLo; + + /* First do InvShiftRows on the State. */ +# define XX(_ix) res->w8[_ix] + { UChar old13 = XX(13); + XX(13) = XX(9); XX(9) = XX(5); XX(5) = XX(1); XX(1) = old13; + } + { UChar old14 = XX(14); UChar old10 = XX(10); + XX(14) = XX(6); XX(10) = XX(2); XX(6) = old14; XX(2) = old10; + } + { UChar old3 = XX(3); + XX(3) = XX(7); XX(7) = XX(11); XX(11) = XX(15); XX(15) = old3; + } +# undef XX + +/* Then do InvSubBytes on the State. */ + UInt i; + for (i = 0; i < 16; i++) { + res->w8[i] = aesMapInvSubBytes[res->w8[i] & 0xFF]; + } +} + + +/* CALLED FROM GENERATED CODE */ +void arm64g_dirtyhelper_AESMC ( /*OUT*/V128* res, ULong argHi, ULong argLo ) +{ + res->w64[1] = argHi; + res->w64[0] = argLo; + aesMixColumn(&res->w8[0]); + aesMixColumn(&res->w8[4]); + aesMixColumn(&res->w8[8]); + aesMixColumn(&res->w8[12]); +} + + +/* CALLED FROM GENERATED CODE */ +void arm64g_dirtyhelper_AESIMC ( /*OUT*/V128* res, ULong argHi, ULong argLo ) +{ + res->w64[1] = argHi; + res->w64[0] = argLo; + aesInvMixColumn(&res->w8[0]); + aesInvMixColumn(&res->w8[4]); + aesInvMixColumn(&res->w8[8]); + aesInvMixColumn(&res->w8[12]); +} + + /*---------------------------------------------------------------*/ /*--- Flag-helpers translation-time function specialisers. ---*/ /*--- These help iropt specialise calls the above run-time ---*/ diff --git a/VEX/priv/guest_arm64_toIR.c b/VEX/priv/guest_arm64_toIR.c index d4fe1b8999..233ae64ffe 100644 --- a/VEX/priv/guest_arm64_toIR.c +++ b/VEX/priv/guest_arm64_toIR.c @@ -12860,7 +12860,75 @@ Bool dis_AdvSIMD_vector_x_indexed_elem(/*MB_OUT*/DisResult* dres, UInt insn) static Bool dis_AdvSIMD_crypto_aes(/*MB_OUT*/DisResult* dres, UInt insn) { + /* 31 23 21 16 11 9 4 + 0100 1110 size 10100 opcode 10 n d + Decode fields are: size,opcode + Size is always 00 in ARMv8, it appears. + */ # define INSN(_bMax,_bMin) SLICE_UInt(insn, (_bMax), (_bMin)) + if (INSN(31,24) != BITS8(0,1,0,0,1,1,1,0) + || INSN(21,17) != BITS5(1,0,1,0,0) || INSN(11,10) != BITS2(1,0)) { + return False; + } + UInt size = INSN(23,22); + UInt opcode = INSN(16,12); + UInt nn = INSN(9,5); + UInt dd = INSN(4,0); + + if (size == BITS2(0,0) + && (opcode == BITS5(0,0,1,0,0) || opcode == BITS5(0,0,1,0,1))) { + /* -------- 00,00100: AESE Vd.16b, Vn.16b -------- */ + /* -------- 00,00101: AESD Vd.16b, Vn.16b -------- */ + Bool isD = opcode == BITS5(0,0,1,0,1); + IRTemp op1 = newTemp(Ity_V128); + IRTemp op2 = newTemp(Ity_V128); + IRTemp xord = newTemp(Ity_V128); + IRTemp res = newTemp(Ity_V128); + void* helper = isD ? &arm64g_dirtyhelper_AESD + : &arm64g_dirtyhelper_AESE; + const HChar* hname = isD ? "arm64g_dirtyhelper_AESD" + : "arm64g_dirtyhelper_AESE"; + assign(op1, getQReg128(dd)); + assign(op2, getQReg128(nn)); + assign(xord, binop(Iop_XorV128, mkexpr(op1), mkexpr(op2))); + IRDirty* di + = unsafeIRDirty_1_N( res, 0/*regparms*/, hname, helper, + mkIRExprVec_3( + IRExpr_VECRET(), + unop(Iop_V128HIto64, mkexpr(xord)), + unop(Iop_V128to64, mkexpr(xord)) ) ); + stmt(IRStmt_Dirty(di)); + putQReg128(dd, mkexpr(res)); + DIP("aes%c %s.16b, %s.16b\n", isD ? 'd' : 'e', + nameQReg128(dd), nameQReg128(nn)); + return True; + } + + if (size == BITS2(0,0) + && (opcode == BITS5(0,0,1,1,0) || opcode == BITS5(0,0,1,1,1))) { + /* -------- 00,00110: AESMC Vd.16b, Vn.16b -------- */ + /* -------- 00,00111: AESIMC Vd.16b, Vn.16b -------- */ + Bool isI = opcode == BITS5(0,0,1,1,1); + IRTemp src = newTemp(Ity_V128); + IRTemp res = newTemp(Ity_V128); + void* helper = isI ? &arm64g_dirtyhelper_AESIMC + : &arm64g_dirtyhelper_AESMC; + const HChar* hname = isI ? "arm64g_dirtyhelper_AESIMC" + : "arm64g_dirtyhelper_AESMC"; + assign(src, getQReg128(nn)); + IRDirty* di + = unsafeIRDirty_1_N( res, 0/*regparms*/, hname, helper, + mkIRExprVec_3( + IRExpr_VECRET(), + unop(Iop_V128HIto64, mkexpr(src)), + unop(Iop_V128to64, mkexpr(src)) ) ); + stmt(IRStmt_Dirty(di)); + putQReg128(dd, mkexpr(res)); + DIP("aes%s %s.16b, %s.16b\n", isI ? "imc" : "mc", + nameQReg128(dd), nameQReg128(nn)); + return True; + } + return False; # undef INSN }