ring-native 0.0.0 → 0.1.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/.gitignore +1 -0
- data/CHANGES.md +7 -0
- data/Makefile +5 -0
- data/README.md +12 -5
- data/Rakefile +4 -0
- data/ext/ring/extconf.rb +4 -5
- data/lib/ring/native.rb +3 -1
- data/lib/ring/native/version.rb +5 -1
- data/ring-native.gemspec +6 -6
- data/vendor/ring-ffi/Cargo.lock +26 -0
- data/vendor/ring-ffi/Cargo.toml +45 -0
- data/vendor/ring-ffi/LICENSE +16 -0
- data/vendor/ring-ffi/README.md +59 -0
- data/vendor/ring-ffi/src/lib.rs +79 -0
- metadata +10 -255
- data/vendor/ring/BUILDING.md +0 -40
- data/vendor/ring/Cargo.toml +0 -43
- data/vendor/ring/LICENSE +0 -185
- data/vendor/ring/Makefile +0 -35
- data/vendor/ring/PORTING.md +0 -163
- data/vendor/ring/README.md +0 -113
- data/vendor/ring/STYLE.md +0 -197
- data/vendor/ring/appveyor.yml +0 -27
- data/vendor/ring/build.rs +0 -108
- data/vendor/ring/crypto/aes/aes.c +0 -1142
- data/vendor/ring/crypto/aes/aes_test.Windows.vcxproj +0 -25
- data/vendor/ring/crypto/aes/aes_test.cc +0 -93
- data/vendor/ring/crypto/aes/asm/aes-586.pl +0 -2368
- data/vendor/ring/crypto/aes/asm/aes-armv4.pl +0 -1249
- data/vendor/ring/crypto/aes/asm/aes-x86_64.pl +0 -2246
- data/vendor/ring/crypto/aes/asm/aesni-x86.pl +0 -1318
- data/vendor/ring/crypto/aes/asm/aesni-x86_64.pl +0 -2084
- data/vendor/ring/crypto/aes/asm/aesv8-armx.pl +0 -675
- data/vendor/ring/crypto/aes/asm/bsaes-armv7.pl +0 -1364
- data/vendor/ring/crypto/aes/asm/bsaes-x86_64.pl +0 -1565
- data/vendor/ring/crypto/aes/asm/vpaes-x86.pl +0 -841
- data/vendor/ring/crypto/aes/asm/vpaes-x86_64.pl +0 -1116
- data/vendor/ring/crypto/aes/internal.h +0 -87
- data/vendor/ring/crypto/aes/mode_wrappers.c +0 -61
- data/vendor/ring/crypto/bn/add.c +0 -394
- data/vendor/ring/crypto/bn/asm/armv4-mont.pl +0 -694
- data/vendor/ring/crypto/bn/asm/armv8-mont.pl +0 -1503
- data/vendor/ring/crypto/bn/asm/bn-586.pl +0 -774
- data/vendor/ring/crypto/bn/asm/co-586.pl +0 -287
- data/vendor/ring/crypto/bn/asm/rsaz-avx2.pl +0 -1882
- data/vendor/ring/crypto/bn/asm/x86-mont.pl +0 -592
- data/vendor/ring/crypto/bn/asm/x86_64-gcc.c +0 -599
- data/vendor/ring/crypto/bn/asm/x86_64-mont.pl +0 -1393
- data/vendor/ring/crypto/bn/asm/x86_64-mont5.pl +0 -3507
- data/vendor/ring/crypto/bn/bn.c +0 -352
- data/vendor/ring/crypto/bn/bn_asn1.c +0 -74
- data/vendor/ring/crypto/bn/bn_test.Windows.vcxproj +0 -25
- data/vendor/ring/crypto/bn/bn_test.cc +0 -1696
- data/vendor/ring/crypto/bn/cmp.c +0 -200
- data/vendor/ring/crypto/bn/convert.c +0 -433
- data/vendor/ring/crypto/bn/ctx.c +0 -311
- data/vendor/ring/crypto/bn/div.c +0 -594
- data/vendor/ring/crypto/bn/exponentiation.c +0 -1335
- data/vendor/ring/crypto/bn/gcd.c +0 -711
- data/vendor/ring/crypto/bn/generic.c +0 -1019
- data/vendor/ring/crypto/bn/internal.h +0 -316
- data/vendor/ring/crypto/bn/montgomery.c +0 -516
- data/vendor/ring/crypto/bn/mul.c +0 -888
- data/vendor/ring/crypto/bn/prime.c +0 -829
- data/vendor/ring/crypto/bn/random.c +0 -334
- data/vendor/ring/crypto/bn/rsaz_exp.c +0 -262
- data/vendor/ring/crypto/bn/rsaz_exp.h +0 -53
- data/vendor/ring/crypto/bn/shift.c +0 -276
- data/vendor/ring/crypto/bytestring/bytestring_test.Windows.vcxproj +0 -25
- data/vendor/ring/crypto/bytestring/bytestring_test.cc +0 -421
- data/vendor/ring/crypto/bytestring/cbb.c +0 -399
- data/vendor/ring/crypto/bytestring/cbs.c +0 -227
- data/vendor/ring/crypto/bytestring/internal.h +0 -46
- data/vendor/ring/crypto/chacha/chacha_generic.c +0 -140
- data/vendor/ring/crypto/chacha/chacha_vec.c +0 -323
- data/vendor/ring/crypto/chacha/chacha_vec_arm.S +0 -1447
- data/vendor/ring/crypto/chacha/chacha_vec_arm_generate.go +0 -153
- data/vendor/ring/crypto/cipher/cipher_test.Windows.vcxproj +0 -25
- data/vendor/ring/crypto/cipher/e_aes.c +0 -390
- data/vendor/ring/crypto/cipher/e_chacha20poly1305.c +0 -208
- data/vendor/ring/crypto/cipher/internal.h +0 -173
- data/vendor/ring/crypto/cipher/test/aes_128_gcm_tests.txt +0 -543
- data/vendor/ring/crypto/cipher/test/aes_128_key_wrap_tests.txt +0 -9
- data/vendor/ring/crypto/cipher/test/aes_256_gcm_tests.txt +0 -475
- data/vendor/ring/crypto/cipher/test/aes_256_key_wrap_tests.txt +0 -23
- data/vendor/ring/crypto/cipher/test/chacha20_poly1305_old_tests.txt +0 -422
- data/vendor/ring/crypto/cipher/test/chacha20_poly1305_tests.txt +0 -484
- data/vendor/ring/crypto/cipher/test/cipher_test.txt +0 -100
- data/vendor/ring/crypto/constant_time_test.Windows.vcxproj +0 -25
- data/vendor/ring/crypto/constant_time_test.c +0 -304
- data/vendor/ring/crypto/cpu-arm-asm.S +0 -32
- data/vendor/ring/crypto/cpu-arm.c +0 -199
- data/vendor/ring/crypto/cpu-intel.c +0 -261
- data/vendor/ring/crypto/crypto.c +0 -151
- data/vendor/ring/crypto/curve25519/asm/x25519-arm.S +0 -2118
- data/vendor/ring/crypto/curve25519/curve25519.c +0 -4888
- data/vendor/ring/crypto/curve25519/x25519_test.cc +0 -128
- data/vendor/ring/crypto/digest/md32_common.h +0 -181
- data/vendor/ring/crypto/ec/asm/p256-x86_64-asm.pl +0 -2725
- data/vendor/ring/crypto/ec/ec.c +0 -193
- data/vendor/ring/crypto/ec/ec_curves.c +0 -61
- data/vendor/ring/crypto/ec/ec_key.c +0 -228
- data/vendor/ring/crypto/ec/ec_montgomery.c +0 -114
- data/vendor/ring/crypto/ec/example_mul.Windows.vcxproj +0 -25
- data/vendor/ring/crypto/ec/internal.h +0 -243
- data/vendor/ring/crypto/ec/oct.c +0 -253
- data/vendor/ring/crypto/ec/p256-64.c +0 -1794
- data/vendor/ring/crypto/ec/p256-x86_64-table.h +0 -9548
- data/vendor/ring/crypto/ec/p256-x86_64.c +0 -509
- data/vendor/ring/crypto/ec/simple.c +0 -1007
- data/vendor/ring/crypto/ec/util-64.c +0 -183
- data/vendor/ring/crypto/ec/wnaf.c +0 -508
- data/vendor/ring/crypto/ecdh/ecdh.c +0 -155
- data/vendor/ring/crypto/ecdsa/ecdsa.c +0 -304
- data/vendor/ring/crypto/ecdsa/ecdsa_asn1.c +0 -193
- data/vendor/ring/crypto/ecdsa/ecdsa_test.Windows.vcxproj +0 -25
- data/vendor/ring/crypto/ecdsa/ecdsa_test.cc +0 -327
- data/vendor/ring/crypto/header_removed.h +0 -17
- data/vendor/ring/crypto/internal.h +0 -495
- data/vendor/ring/crypto/libring.Windows.vcxproj +0 -101
- data/vendor/ring/crypto/mem.c +0 -98
- data/vendor/ring/crypto/modes/asm/aesni-gcm-x86_64.pl +0 -1045
- data/vendor/ring/crypto/modes/asm/ghash-armv4.pl +0 -517
- data/vendor/ring/crypto/modes/asm/ghash-x86.pl +0 -1393
- data/vendor/ring/crypto/modes/asm/ghash-x86_64.pl +0 -1741
- data/vendor/ring/crypto/modes/asm/ghashv8-armx.pl +0 -422
- data/vendor/ring/crypto/modes/ctr.c +0 -226
- data/vendor/ring/crypto/modes/gcm.c +0 -1206
- data/vendor/ring/crypto/modes/gcm_test.Windows.vcxproj +0 -25
- data/vendor/ring/crypto/modes/gcm_test.c +0 -348
- data/vendor/ring/crypto/modes/internal.h +0 -299
- data/vendor/ring/crypto/perlasm/arm-xlate.pl +0 -170
- data/vendor/ring/crypto/perlasm/readme +0 -100
- data/vendor/ring/crypto/perlasm/x86_64-xlate.pl +0 -1164
- data/vendor/ring/crypto/perlasm/x86asm.pl +0 -292
- data/vendor/ring/crypto/perlasm/x86gas.pl +0 -263
- data/vendor/ring/crypto/perlasm/x86masm.pl +0 -200
- data/vendor/ring/crypto/perlasm/x86nasm.pl +0 -187
- data/vendor/ring/crypto/poly1305/poly1305.c +0 -331
- data/vendor/ring/crypto/poly1305/poly1305_arm.c +0 -301
- data/vendor/ring/crypto/poly1305/poly1305_arm_asm.S +0 -2015
- data/vendor/ring/crypto/poly1305/poly1305_test.Windows.vcxproj +0 -25
- data/vendor/ring/crypto/poly1305/poly1305_test.cc +0 -80
- data/vendor/ring/crypto/poly1305/poly1305_test.txt +0 -52
- data/vendor/ring/crypto/poly1305/poly1305_vec.c +0 -892
- data/vendor/ring/crypto/rand/asm/rdrand-x86_64.pl +0 -75
- data/vendor/ring/crypto/rand/internal.h +0 -32
- data/vendor/ring/crypto/rand/rand.c +0 -189
- data/vendor/ring/crypto/rand/urandom.c +0 -219
- data/vendor/ring/crypto/rand/windows.c +0 -56
- data/vendor/ring/crypto/refcount_c11.c +0 -66
- data/vendor/ring/crypto/refcount_lock.c +0 -53
- data/vendor/ring/crypto/refcount_test.Windows.vcxproj +0 -25
- data/vendor/ring/crypto/refcount_test.c +0 -58
- data/vendor/ring/crypto/rsa/blinding.c +0 -462
- data/vendor/ring/crypto/rsa/internal.h +0 -108
- data/vendor/ring/crypto/rsa/padding.c +0 -300
- data/vendor/ring/crypto/rsa/rsa.c +0 -450
- data/vendor/ring/crypto/rsa/rsa_asn1.c +0 -261
- data/vendor/ring/crypto/rsa/rsa_impl.c +0 -944
- data/vendor/ring/crypto/rsa/rsa_test.Windows.vcxproj +0 -25
- data/vendor/ring/crypto/rsa/rsa_test.cc +0 -437
- data/vendor/ring/crypto/sha/asm/sha-armv8.pl +0 -436
- data/vendor/ring/crypto/sha/asm/sha-x86_64.pl +0 -2390
- data/vendor/ring/crypto/sha/asm/sha256-586.pl +0 -1275
- data/vendor/ring/crypto/sha/asm/sha256-armv4.pl +0 -735
- data/vendor/ring/crypto/sha/asm/sha256-armv8.pl +0 -14
- data/vendor/ring/crypto/sha/asm/sha256-x86_64.pl +0 -14
- data/vendor/ring/crypto/sha/asm/sha512-586.pl +0 -911
- data/vendor/ring/crypto/sha/asm/sha512-armv4.pl +0 -666
- data/vendor/ring/crypto/sha/asm/sha512-armv8.pl +0 -14
- data/vendor/ring/crypto/sha/asm/sha512-x86_64.pl +0 -14
- data/vendor/ring/crypto/sha/sha1.c +0 -271
- data/vendor/ring/crypto/sha/sha256.c +0 -204
- data/vendor/ring/crypto/sha/sha512.c +0 -355
- data/vendor/ring/crypto/test/file_test.cc +0 -326
- data/vendor/ring/crypto/test/file_test.h +0 -181
- data/vendor/ring/crypto/test/malloc.cc +0 -150
- data/vendor/ring/crypto/test/scoped_types.h +0 -95
- data/vendor/ring/crypto/test/test.Windows.vcxproj +0 -35
- data/vendor/ring/crypto/test/test_util.cc +0 -46
- data/vendor/ring/crypto/test/test_util.h +0 -41
- data/vendor/ring/crypto/thread_none.c +0 -55
- data/vendor/ring/crypto/thread_pthread.c +0 -165
- data/vendor/ring/crypto/thread_test.Windows.vcxproj +0 -25
- data/vendor/ring/crypto/thread_test.c +0 -200
- data/vendor/ring/crypto/thread_win.c +0 -282
- data/vendor/ring/examples/checkdigest.rs +0 -103
- data/vendor/ring/include/openssl/aes.h +0 -121
- data/vendor/ring/include/openssl/arm_arch.h +0 -129
- data/vendor/ring/include/openssl/base.h +0 -156
- data/vendor/ring/include/openssl/bn.h +0 -794
- data/vendor/ring/include/openssl/buffer.h +0 -18
- data/vendor/ring/include/openssl/bytestring.h +0 -235
- data/vendor/ring/include/openssl/chacha.h +0 -37
- data/vendor/ring/include/openssl/cmac.h +0 -76
- data/vendor/ring/include/openssl/cpu.h +0 -184
- data/vendor/ring/include/openssl/crypto.h +0 -43
- data/vendor/ring/include/openssl/curve25519.h +0 -88
- data/vendor/ring/include/openssl/ec.h +0 -225
- data/vendor/ring/include/openssl/ec_key.h +0 -129
- data/vendor/ring/include/openssl/ecdh.h +0 -110
- data/vendor/ring/include/openssl/ecdsa.h +0 -156
- data/vendor/ring/include/openssl/err.h +0 -201
- data/vendor/ring/include/openssl/mem.h +0 -101
- data/vendor/ring/include/openssl/obj_mac.h +0 -71
- data/vendor/ring/include/openssl/opensslfeatures.h +0 -68
- data/vendor/ring/include/openssl/opensslv.h +0 -18
- data/vendor/ring/include/openssl/ossl_typ.h +0 -18
- data/vendor/ring/include/openssl/poly1305.h +0 -51
- data/vendor/ring/include/openssl/rand.h +0 -70
- data/vendor/ring/include/openssl/rsa.h +0 -399
- data/vendor/ring/include/openssl/thread.h +0 -133
- data/vendor/ring/include/openssl/type_check.h +0 -71
- data/vendor/ring/mk/Common.props +0 -63
- data/vendor/ring/mk/Windows.props +0 -42
- data/vendor/ring/mk/WindowsTest.props +0 -18
- data/vendor/ring/mk/appveyor.bat +0 -62
- data/vendor/ring/mk/bottom_of_makefile.mk +0 -54
- data/vendor/ring/mk/ring.mk +0 -266
- data/vendor/ring/mk/top_of_makefile.mk +0 -214
- data/vendor/ring/mk/travis.sh +0 -40
- data/vendor/ring/mk/update-travis-yml.py +0 -229
- data/vendor/ring/ring.sln +0 -153
- data/vendor/ring/src/aead.rs +0 -682
- data/vendor/ring/src/agreement.rs +0 -248
- data/vendor/ring/src/c.rs +0 -129
- data/vendor/ring/src/constant_time.rs +0 -37
- data/vendor/ring/src/der.rs +0 -96
- data/vendor/ring/src/digest.rs +0 -690
- data/vendor/ring/src/digest_tests.txt +0 -57
- data/vendor/ring/src/ecc.rs +0 -28
- data/vendor/ring/src/ecc_build.rs +0 -279
- data/vendor/ring/src/ecc_curves.rs +0 -117
- data/vendor/ring/src/ed25519_tests.txt +0 -2579
- data/vendor/ring/src/exe_tests.rs +0 -46
- data/vendor/ring/src/ffi.rs +0 -29
- data/vendor/ring/src/file_test.rs +0 -187
- data/vendor/ring/src/hkdf.rs +0 -153
- data/vendor/ring/src/hkdf_tests.txt +0 -59
- data/vendor/ring/src/hmac.rs +0 -414
- data/vendor/ring/src/hmac_tests.txt +0 -97
- data/vendor/ring/src/input.rs +0 -312
- data/vendor/ring/src/lib.rs +0 -41
- data/vendor/ring/src/pbkdf2.rs +0 -265
- data/vendor/ring/src/pbkdf2_tests.txt +0 -113
- data/vendor/ring/src/polyfill.rs +0 -57
- data/vendor/ring/src/rand.rs +0 -28
- data/vendor/ring/src/signature.rs +0 -314
- data/vendor/ring/third-party/NIST/README.md +0 -9
- data/vendor/ring/third-party/NIST/SHAVS/SHA1LongMsg.rsp +0 -263
- data/vendor/ring/third-party/NIST/SHAVS/SHA1Monte.rsp +0 -309
- data/vendor/ring/third-party/NIST/SHAVS/SHA1ShortMsg.rsp +0 -267
- data/vendor/ring/third-party/NIST/SHAVS/SHA224LongMsg.rsp +0 -263
- data/vendor/ring/third-party/NIST/SHAVS/SHA224Monte.rsp +0 -309
- data/vendor/ring/third-party/NIST/SHAVS/SHA224ShortMsg.rsp +0 -267
- data/vendor/ring/third-party/NIST/SHAVS/SHA256LongMsg.rsp +0 -263
- data/vendor/ring/third-party/NIST/SHAVS/SHA256Monte.rsp +0 -309
- data/vendor/ring/third-party/NIST/SHAVS/SHA256ShortMsg.rsp +0 -267
- data/vendor/ring/third-party/NIST/SHAVS/SHA384LongMsg.rsp +0 -519
- data/vendor/ring/third-party/NIST/SHAVS/SHA384Monte.rsp +0 -309
- data/vendor/ring/third-party/NIST/SHAVS/SHA384ShortMsg.rsp +0 -523
- data/vendor/ring/third-party/NIST/SHAVS/SHA512LongMsg.rsp +0 -519
- data/vendor/ring/third-party/NIST/SHAVS/SHA512Monte.rsp +0 -309
- data/vendor/ring/third-party/NIST/SHAVS/SHA512ShortMsg.rsp +0 -523
- data/vendor/ring/third-party/NIST/sha256sums.txt +0 -1
@@ -1,1565 +0,0 @@
|
|
1
|
-
#!/usr/bin/env perl
|
2
|
-
|
3
|
-
###################################################################
|
4
|
-
### AES-128 [originally in CTR mode] ###
|
5
|
-
### bitsliced implementation for Intel Core 2 processors ###
|
6
|
-
### requires support of SSE extensions up to SSSE3 ###
|
7
|
-
### Author: Emilia Käsper and Peter Schwabe ###
|
8
|
-
### Date: 2009-03-19 ###
|
9
|
-
### Public domain ###
|
10
|
-
### ###
|
11
|
-
### See http://homes.esat.kuleuven.be/~ekasper/#software for ###
|
12
|
-
### further information. ###
|
13
|
-
###################################################################
|
14
|
-
#
|
15
|
-
# September 2011.
|
16
|
-
#
|
17
|
-
# Started as transliteration to "perlasm" the original code has
|
18
|
-
# undergone following changes:
|
19
|
-
#
|
20
|
-
# - code was made position-independent;
|
21
|
-
# - rounds were folded into a loop resulting in >5x size reduction
|
22
|
-
# from 12.5KB to 2.2KB;
|
23
|
-
# - above was possibile thanks to mixcolumns() modification that
|
24
|
-
# allowed to feed its output back to aesenc[last], this was
|
25
|
-
# achieved at cost of two additional inter-registers moves;
|
26
|
-
# - some instruction reordering and interleaving;
|
27
|
-
# - this module doesn't implement key setup subroutine, instead it
|
28
|
-
# relies on conversion of "conventional" key schedule as returned
|
29
|
-
# by AES_set_encrypt_key (see discussion below);
|
30
|
-
# - first and last round keys are treated differently, which allowed
|
31
|
-
# to skip one shiftrows(), reduce bit-sliced key schedule and
|
32
|
-
# speed-up conversion by 22%;
|
33
|
-
# - support for 192- and 256-bit keys was added;
|
34
|
-
#
|
35
|
-
# Resulting performance in CPU cycles spent to encrypt one byte out
|
36
|
-
# of 4096-byte buffer with 128-bit key is:
|
37
|
-
#
|
38
|
-
# Emilia's this(*) difference
|
39
|
-
#
|
40
|
-
# Core 2 9.30 8.69 +7%
|
41
|
-
# Nehalem(**) 7.63 6.88 +11%
|
42
|
-
# Atom 17.1 16.4 +4%
|
43
|
-
# Silvermont - 12.9
|
44
|
-
#
|
45
|
-
# (*) Comparison is not completely fair, because "this" is ECB,
|
46
|
-
# i.e. no extra processing such as counter values calculation
|
47
|
-
# and xor-ing input as in Emilia's CTR implementation is
|
48
|
-
# performed. However, the CTR calculations stand for not more
|
49
|
-
# than 1% of total time, so comparison is *rather* fair.
|
50
|
-
#
|
51
|
-
# (**) Results were collected on Westmere, which is considered to
|
52
|
-
# be equivalent to Nehalem for this code.
|
53
|
-
#
|
54
|
-
# As for key schedule conversion subroutine. Interface to OpenSSL
|
55
|
-
# relies on per-invocation on-the-fly conversion. This naturally
|
56
|
-
# has impact on performance, especially for short inputs. Conversion
|
57
|
-
# time in CPU cycles and its ratio to CPU cycles spent in 8x block
|
58
|
-
# function is:
|
59
|
-
#
|
60
|
-
# conversion conversion/8x block
|
61
|
-
# Core 2 240 0.22
|
62
|
-
# Nehalem 180 0.20
|
63
|
-
# Atom 430 0.20
|
64
|
-
#
|
65
|
-
# The ratio values mean that 128-byte blocks will be processed
|
66
|
-
# 16-18% slower, 256-byte blocks - 9-10%, 384-byte blocks - 6-7%,
|
67
|
-
# etc. Then keep in mind that input sizes not divisible by 128 are
|
68
|
-
# *effectively* slower, especially shortest ones, e.g. consecutive
|
69
|
-
# 144-byte blocks are processed 44% slower than one would expect,
|
70
|
-
# 272 - 29%, 400 - 22%, etc. Yet, despite all these "shortcomings"
|
71
|
-
# it's still faster than ["hyper-threading-safe" code path in]
|
72
|
-
# aes-x86_64.pl on all lengths above 64 bytes...
|
73
|
-
#
|
74
|
-
# October 2011.
|
75
|
-
#
|
76
|
-
# Add decryption procedure. Performance in CPU cycles spent to decrypt
|
77
|
-
# one byte out of 4096-byte buffer with 128-bit key is:
|
78
|
-
#
|
79
|
-
# Core 2 9.98
|
80
|
-
# Nehalem 7.80
|
81
|
-
# Atom 17.9
|
82
|
-
# Silvermont 14.0
|
83
|
-
#
|
84
|
-
# November 2011.
|
85
|
-
#
|
86
|
-
# Add bsaes_xts_[en|de]crypt. Less-than-80-bytes-block performance is
|
87
|
-
# suboptimal, but XTS is meant to be used with larger blocks...
|
88
|
-
#
|
89
|
-
# <appro@openssl.org>
|
90
|
-
|
91
|
-
$flavour = shift;
|
92
|
-
$output = shift;
|
93
|
-
if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
|
94
|
-
|
95
|
-
$win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
|
96
|
-
|
97
|
-
$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
|
98
|
-
( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
|
99
|
-
( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
|
100
|
-
die "can't locate x86_64-xlate.pl";
|
101
|
-
|
102
|
-
open OUT,"| \"$^X\" $xlate $flavour $output";
|
103
|
-
*STDOUT=*OUT;
|
104
|
-
|
105
|
-
my ($inp,$out,$len,$key,$ivp)=("%rdi","%rsi","%rdx","%rcx");
|
106
|
-
my @XMM=map("%xmm$_",(15,0..14)); # best on Atom, +10% over (0..15)
|
107
|
-
|
108
|
-
{
|
109
|
-
my ($key,$rounds,$const)=("%rax","%r10d","%r11");
|
110
|
-
|
111
|
-
sub Sbox {
|
112
|
-
# input in lsb > [b0, b1, b2, b3, b4, b5, b6, b7] < msb
|
113
|
-
# output in lsb > [b0, b1, b4, b6, b3, b7, b2, b5] < msb
|
114
|
-
my @b=@_[0..7];
|
115
|
-
my @t=@_[8..11];
|
116
|
-
my @s=@_[12..15];
|
117
|
-
&InBasisChange (@b);
|
118
|
-
&Inv_GF256 (@b[6,5,0,3,7,1,4,2],@t,@s);
|
119
|
-
&OutBasisChange (@b[7,1,4,2,6,5,0,3]);
|
120
|
-
}
|
121
|
-
|
122
|
-
sub InBasisChange {
|
123
|
-
# input in lsb > [b0, b1, b2, b3, b4, b5, b6, b7] < msb
|
124
|
-
# output in lsb > [b6, b5, b0, b3, b7, b1, b4, b2] < msb
|
125
|
-
my @b=@_[0..7];
|
126
|
-
$code.=<<___;
|
127
|
-
pxor @b[6], @b[5]
|
128
|
-
pxor @b[1], @b[2]
|
129
|
-
pxor @b[0], @b[3]
|
130
|
-
pxor @b[2], @b[6]
|
131
|
-
pxor @b[0], @b[5]
|
132
|
-
|
133
|
-
pxor @b[3], @b[6]
|
134
|
-
pxor @b[7], @b[3]
|
135
|
-
pxor @b[5], @b[7]
|
136
|
-
pxor @b[4], @b[3]
|
137
|
-
pxor @b[5], @b[4]
|
138
|
-
pxor @b[1], @b[3]
|
139
|
-
|
140
|
-
pxor @b[7], @b[2]
|
141
|
-
pxor @b[5], @b[1]
|
142
|
-
___
|
143
|
-
}
|
144
|
-
|
145
|
-
sub OutBasisChange {
|
146
|
-
# input in lsb > [b0, b1, b2, b3, b4, b5, b6, b7] < msb
|
147
|
-
# output in lsb > [b6, b1, b2, b4, b7, b0, b3, b5] < msb
|
148
|
-
my @b=@_[0..7];
|
149
|
-
$code.=<<___;
|
150
|
-
pxor @b[6], @b[0]
|
151
|
-
pxor @b[4], @b[1]
|
152
|
-
pxor @b[0], @b[2]
|
153
|
-
pxor @b[6], @b[4]
|
154
|
-
pxor @b[1], @b[6]
|
155
|
-
|
156
|
-
pxor @b[5], @b[1]
|
157
|
-
pxor @b[3], @b[5]
|
158
|
-
pxor @b[7], @b[3]
|
159
|
-
pxor @b[5], @b[7]
|
160
|
-
pxor @b[5], @b[2]
|
161
|
-
|
162
|
-
pxor @b[7], @b[4]
|
163
|
-
___
|
164
|
-
}
|
165
|
-
|
166
|
-
sub InvSbox {
|
167
|
-
# input in lsb > [b0, b1, b2, b3, b4, b5, b6, b7] < msb
|
168
|
-
# output in lsb > [b0, b1, b6, b4, b2, b7, b3, b5] < msb
|
169
|
-
my @b=@_[0..7];
|
170
|
-
my @t=@_[8..11];
|
171
|
-
my @s=@_[12..15];
|
172
|
-
&InvInBasisChange (@b);
|
173
|
-
&Inv_GF256 (@b[5,1,2,6,3,7,0,4],@t,@s);
|
174
|
-
&InvOutBasisChange (@b[3,7,0,4,5,1,2,6]);
|
175
|
-
}
|
176
|
-
|
177
|
-
sub InvInBasisChange { # OutBasisChange in reverse
|
178
|
-
my @b=@_[5,1,2,6,3,7,0,4];
|
179
|
-
$code.=<<___
|
180
|
-
pxor @b[7], @b[4]
|
181
|
-
|
182
|
-
pxor @b[5], @b[7]
|
183
|
-
pxor @b[5], @b[2]
|
184
|
-
pxor @b[7], @b[3]
|
185
|
-
pxor @b[3], @b[5]
|
186
|
-
pxor @b[5], @b[1]
|
187
|
-
|
188
|
-
pxor @b[1], @b[6]
|
189
|
-
pxor @b[0], @b[2]
|
190
|
-
pxor @b[6], @b[4]
|
191
|
-
pxor @b[6], @b[0]
|
192
|
-
pxor @b[4], @b[1]
|
193
|
-
___
|
194
|
-
}
|
195
|
-
|
196
|
-
sub InvOutBasisChange { # InBasisChange in reverse
|
197
|
-
my @b=@_[2,5,7,3,6,1,0,4];
|
198
|
-
$code.=<<___;
|
199
|
-
pxor @b[5], @b[1]
|
200
|
-
pxor @b[7], @b[2]
|
201
|
-
|
202
|
-
pxor @b[1], @b[3]
|
203
|
-
pxor @b[5], @b[4]
|
204
|
-
pxor @b[5], @b[7]
|
205
|
-
pxor @b[4], @b[3]
|
206
|
-
pxor @b[0], @b[5]
|
207
|
-
pxor @b[7], @b[3]
|
208
|
-
pxor @b[2], @b[6]
|
209
|
-
pxor @b[1], @b[2]
|
210
|
-
pxor @b[3], @b[6]
|
211
|
-
|
212
|
-
pxor @b[0], @b[3]
|
213
|
-
pxor @b[6], @b[5]
|
214
|
-
___
|
215
|
-
}
|
216
|
-
|
217
|
-
sub Mul_GF4 {
|
218
|
-
#;*************************************************************
|
219
|
-
#;* Mul_GF4: Input x0-x1,y0-y1 Output x0-x1 Temp t0 (8) *
|
220
|
-
#;*************************************************************
|
221
|
-
my ($x0,$x1,$y0,$y1,$t0)=@_;
|
222
|
-
$code.=<<___;
|
223
|
-
movdqa $y0, $t0
|
224
|
-
pxor $y1, $t0
|
225
|
-
pand $x0, $t0
|
226
|
-
pxor $x1, $x0
|
227
|
-
pand $y0, $x1
|
228
|
-
pand $y1, $x0
|
229
|
-
pxor $x1, $x0
|
230
|
-
pxor $t0, $x1
|
231
|
-
___
|
232
|
-
}
|
233
|
-
|
234
|
-
sub Mul_GF4_N { # not used, see next subroutine
|
235
|
-
# multiply and scale by N
|
236
|
-
my ($x0,$x1,$y0,$y1,$t0)=@_;
|
237
|
-
$code.=<<___;
|
238
|
-
movdqa $y0, $t0
|
239
|
-
pxor $y1, $t0
|
240
|
-
pand $x0, $t0
|
241
|
-
pxor $x1, $x0
|
242
|
-
pand $y0, $x1
|
243
|
-
pand $y1, $x0
|
244
|
-
pxor $x0, $x1
|
245
|
-
pxor $t0, $x0
|
246
|
-
___
|
247
|
-
}
|
248
|
-
|
249
|
-
sub Mul_GF4_N_GF4 {
|
250
|
-
# interleaved Mul_GF4_N and Mul_GF4
|
251
|
-
my ($x0,$x1,$y0,$y1,$t0,
|
252
|
-
$x2,$x3,$y2,$y3,$t1)=@_;
|
253
|
-
$code.=<<___;
|
254
|
-
movdqa $y0, $t0
|
255
|
-
movdqa $y2, $t1
|
256
|
-
pxor $y1, $t0
|
257
|
-
pxor $y3, $t1
|
258
|
-
pand $x0, $t0
|
259
|
-
pand $x2, $t1
|
260
|
-
pxor $x1, $x0
|
261
|
-
pxor $x3, $x2
|
262
|
-
pand $y0, $x1
|
263
|
-
pand $y2, $x3
|
264
|
-
pand $y1, $x0
|
265
|
-
pand $y3, $x2
|
266
|
-
pxor $x0, $x1
|
267
|
-
pxor $x3, $x2
|
268
|
-
pxor $t0, $x0
|
269
|
-
pxor $t1, $x3
|
270
|
-
___
|
271
|
-
}
|
272
|
-
sub Mul_GF16_2 {
|
273
|
-
my @x=@_[0..7];
|
274
|
-
my @y=@_[8..11];
|
275
|
-
my @t=@_[12..15];
|
276
|
-
$code.=<<___;
|
277
|
-
movdqa @x[0], @t[0]
|
278
|
-
movdqa @x[1], @t[1]
|
279
|
-
___
|
280
|
-
&Mul_GF4 (@x[0], @x[1], @y[0], @y[1], @t[2]);
|
281
|
-
$code.=<<___;
|
282
|
-
pxor @x[2], @t[0]
|
283
|
-
pxor @x[3], @t[1]
|
284
|
-
pxor @y[2], @y[0]
|
285
|
-
pxor @y[3], @y[1]
|
286
|
-
___
|
287
|
-
Mul_GF4_N_GF4 (@t[0], @t[1], @y[0], @y[1], @t[3],
|
288
|
-
@x[2], @x[3], @y[2], @y[3], @t[2]);
|
289
|
-
$code.=<<___;
|
290
|
-
pxor @t[0], @x[0]
|
291
|
-
pxor @t[0], @x[2]
|
292
|
-
pxor @t[1], @x[1]
|
293
|
-
pxor @t[1], @x[3]
|
294
|
-
|
295
|
-
movdqa @x[4], @t[0]
|
296
|
-
movdqa @x[5], @t[1]
|
297
|
-
pxor @x[6], @t[0]
|
298
|
-
pxor @x[7], @t[1]
|
299
|
-
___
|
300
|
-
&Mul_GF4_N_GF4 (@t[0], @t[1], @y[0], @y[1], @t[3],
|
301
|
-
@x[6], @x[7], @y[2], @y[3], @t[2]);
|
302
|
-
$code.=<<___;
|
303
|
-
pxor @y[2], @y[0]
|
304
|
-
pxor @y[3], @y[1]
|
305
|
-
___
|
306
|
-
&Mul_GF4 (@x[4], @x[5], @y[0], @y[1], @t[3]);
|
307
|
-
$code.=<<___;
|
308
|
-
pxor @t[0], @x[4]
|
309
|
-
pxor @t[0], @x[6]
|
310
|
-
pxor @t[1], @x[5]
|
311
|
-
pxor @t[1], @x[7]
|
312
|
-
___
|
313
|
-
}
|
314
|
-
sub Inv_GF256 {
|
315
|
-
#;********************************************************************
|
316
|
-
#;* Inv_GF256: Input x0-x7 Output x0-x7 Temp t0-t3,s0-s3 (144) *
|
317
|
-
#;********************************************************************
|
318
|
-
my @x=@_[0..7];
|
319
|
-
my @t=@_[8..11];
|
320
|
-
my @s=@_[12..15];
|
321
|
-
# direct optimizations from hardware
|
322
|
-
$code.=<<___;
|
323
|
-
movdqa @x[4], @t[3]
|
324
|
-
movdqa @x[5], @t[2]
|
325
|
-
movdqa @x[1], @t[1]
|
326
|
-
movdqa @x[7], @s[1]
|
327
|
-
movdqa @x[0], @s[0]
|
328
|
-
|
329
|
-
pxor @x[6], @t[3]
|
330
|
-
pxor @x[7], @t[2]
|
331
|
-
pxor @x[3], @t[1]
|
332
|
-
movdqa @t[3], @s[2]
|
333
|
-
pxor @x[6], @s[1]
|
334
|
-
movdqa @t[2], @t[0]
|
335
|
-
pxor @x[2], @s[0]
|
336
|
-
movdqa @t[3], @s[3]
|
337
|
-
|
338
|
-
por @t[1], @t[2]
|
339
|
-
por @s[0], @t[3]
|
340
|
-
pxor @t[0], @s[3]
|
341
|
-
pand @s[0], @s[2]
|
342
|
-
pxor @t[1], @s[0]
|
343
|
-
pand @t[1], @t[0]
|
344
|
-
pand @s[0], @s[3]
|
345
|
-
movdqa @x[3], @s[0]
|
346
|
-
pxor @x[2], @s[0]
|
347
|
-
pand @s[0], @s[1]
|
348
|
-
pxor @s[1], @t[3]
|
349
|
-
pxor @s[1], @t[2]
|
350
|
-
movdqa @x[4], @s[1]
|
351
|
-
movdqa @x[1], @s[0]
|
352
|
-
pxor @x[5], @s[1]
|
353
|
-
pxor @x[0], @s[0]
|
354
|
-
movdqa @s[1], @t[1]
|
355
|
-
pand @s[0], @s[1]
|
356
|
-
por @s[0], @t[1]
|
357
|
-
pxor @s[1], @t[0]
|
358
|
-
pxor @s[3], @t[3]
|
359
|
-
pxor @s[2], @t[2]
|
360
|
-
pxor @s[3], @t[1]
|
361
|
-
movdqa @x[7], @s[0]
|
362
|
-
pxor @s[2], @t[0]
|
363
|
-
movdqa @x[6], @s[1]
|
364
|
-
pxor @s[2], @t[1]
|
365
|
-
movdqa @x[5], @s[2]
|
366
|
-
pand @x[3], @s[0]
|
367
|
-
movdqa @x[4], @s[3]
|
368
|
-
pand @x[2], @s[1]
|
369
|
-
pand @x[1], @s[2]
|
370
|
-
por @x[0], @s[3]
|
371
|
-
pxor @s[0], @t[3]
|
372
|
-
pxor @s[1], @t[2]
|
373
|
-
pxor @s[2], @t[1]
|
374
|
-
pxor @s[3], @t[0]
|
375
|
-
|
376
|
-
#Inv_GF16 \t0, \t1, \t2, \t3, \s0, \s1, \s2, \s3
|
377
|
-
|
378
|
-
# new smaller inversion
|
379
|
-
|
380
|
-
movdqa @t[3], @s[0]
|
381
|
-
pand @t[1], @t[3]
|
382
|
-
pxor @t[2], @s[0]
|
383
|
-
|
384
|
-
movdqa @t[0], @s[2]
|
385
|
-
movdqa @s[0], @s[3]
|
386
|
-
pxor @t[3], @s[2]
|
387
|
-
pand @s[2], @s[3]
|
388
|
-
|
389
|
-
movdqa @t[1], @s[1]
|
390
|
-
pxor @t[2], @s[3]
|
391
|
-
pxor @t[0], @s[1]
|
392
|
-
|
393
|
-
pxor @t[2], @t[3]
|
394
|
-
|
395
|
-
pand @t[3], @s[1]
|
396
|
-
|
397
|
-
movdqa @s[2], @t[2]
|
398
|
-
pxor @t[0], @s[1]
|
399
|
-
|
400
|
-
pxor @s[1], @t[2]
|
401
|
-
pxor @s[1], @t[1]
|
402
|
-
|
403
|
-
pand @t[0], @t[2]
|
404
|
-
|
405
|
-
pxor @t[2], @s[2]
|
406
|
-
pxor @t[2], @t[1]
|
407
|
-
|
408
|
-
pand @s[3], @s[2]
|
409
|
-
|
410
|
-
pxor @s[0], @s[2]
|
411
|
-
___
|
412
|
-
# output in s3, s2, s1, t1
|
413
|
-
|
414
|
-
# Mul_GF16_2 \x0, \x1, \x2, \x3, \x4, \x5, \x6, \x7, \t2, \t3, \t0, \t1, \s0, \s1, \s2, \s3
|
415
|
-
|
416
|
-
# Mul_GF16_2 \x0, \x1, \x2, \x3, \x4, \x5, \x6, \x7, \s3, \s2, \s1, \t1, \s0, \t0, \t2, \t3
|
417
|
-
&Mul_GF16_2(@x,@s[3,2,1],@t[1],@s[0],@t[0,2,3]);
|
418
|
-
|
419
|
-
### output msb > [x3,x2,x1,x0,x7,x6,x5,x4] < lsb
|
420
|
-
}
|
421
|
-
|
422
|
-
# AES linear components
|
423
|
-
|
424
|
-
sub ShiftRows {
|
425
|
-
my @x=@_[0..7];
|
426
|
-
my $mask=pop;
|
427
|
-
$code.=<<___;
|
428
|
-
pxor 0x00($key),@x[0]
|
429
|
-
pxor 0x10($key),@x[1]
|
430
|
-
pxor 0x20($key),@x[2]
|
431
|
-
pxor 0x30($key),@x[3]
|
432
|
-
pshufb $mask,@x[0]
|
433
|
-
pshufb $mask,@x[1]
|
434
|
-
pxor 0x40($key),@x[4]
|
435
|
-
pxor 0x50($key),@x[5]
|
436
|
-
pshufb $mask,@x[2]
|
437
|
-
pshufb $mask,@x[3]
|
438
|
-
pxor 0x60($key),@x[6]
|
439
|
-
pxor 0x70($key),@x[7]
|
440
|
-
pshufb $mask,@x[4]
|
441
|
-
pshufb $mask,@x[5]
|
442
|
-
pshufb $mask,@x[6]
|
443
|
-
pshufb $mask,@x[7]
|
444
|
-
lea 0x80($key),$key
|
445
|
-
___
|
446
|
-
}
|
447
|
-
|
448
|
-
sub MixColumns {
|
449
|
-
# modified to emit output in order suitable for feeding back to aesenc[last]
|
450
|
-
my @x=@_[0..7];
|
451
|
-
my @t=@_[8..15];
|
452
|
-
my $inv=@_[16]; # optional
|
453
|
-
$code.=<<___;
|
454
|
-
pshufd \$0x93, @x[0], @t[0] # x0 <<< 32
|
455
|
-
pshufd \$0x93, @x[1], @t[1]
|
456
|
-
pxor @t[0], @x[0] # x0 ^ (x0 <<< 32)
|
457
|
-
pshufd \$0x93, @x[2], @t[2]
|
458
|
-
pxor @t[1], @x[1]
|
459
|
-
pshufd \$0x93, @x[3], @t[3]
|
460
|
-
pxor @t[2], @x[2]
|
461
|
-
pshufd \$0x93, @x[4], @t[4]
|
462
|
-
pxor @t[3], @x[3]
|
463
|
-
pshufd \$0x93, @x[5], @t[5]
|
464
|
-
pxor @t[4], @x[4]
|
465
|
-
pshufd \$0x93, @x[6], @t[6]
|
466
|
-
pxor @t[5], @x[5]
|
467
|
-
pshufd \$0x93, @x[7], @t[7]
|
468
|
-
pxor @t[6], @x[6]
|
469
|
-
pxor @t[7], @x[7]
|
470
|
-
|
471
|
-
pxor @x[0], @t[1]
|
472
|
-
pxor @x[7], @t[0]
|
473
|
-
pxor @x[7], @t[1]
|
474
|
-
pshufd \$0x4E, @x[0], @x[0] # (x0 ^ (x0 <<< 32)) <<< 64)
|
475
|
-
pxor @x[1], @t[2]
|
476
|
-
pshufd \$0x4E, @x[1], @x[1]
|
477
|
-
pxor @x[4], @t[5]
|
478
|
-
pxor @t[0], @x[0]
|
479
|
-
pxor @x[5], @t[6]
|
480
|
-
pxor @t[1], @x[1]
|
481
|
-
pxor @x[3], @t[4]
|
482
|
-
pshufd \$0x4E, @x[4], @t[0]
|
483
|
-
pxor @x[6], @t[7]
|
484
|
-
pshufd \$0x4E, @x[5], @t[1]
|
485
|
-
pxor @x[2], @t[3]
|
486
|
-
pshufd \$0x4E, @x[3], @x[4]
|
487
|
-
pxor @x[7], @t[3]
|
488
|
-
pshufd \$0x4E, @x[7], @x[5]
|
489
|
-
pxor @x[7], @t[4]
|
490
|
-
pshufd \$0x4E, @x[6], @x[3]
|
491
|
-
pxor @t[4], @t[0]
|
492
|
-
pshufd \$0x4E, @x[2], @x[6]
|
493
|
-
pxor @t[5], @t[1]
|
494
|
-
___
|
495
|
-
$code.=<<___ if (!$inv);
|
496
|
-
pxor @t[3], @x[4]
|
497
|
-
pxor @t[7], @x[5]
|
498
|
-
pxor @t[6], @x[3]
|
499
|
-
movdqa @t[0], @x[2]
|
500
|
-
pxor @t[2], @x[6]
|
501
|
-
movdqa @t[1], @x[7]
|
502
|
-
___
|
503
|
-
$code.=<<___ if ($inv);
|
504
|
-
pxor @x[4], @t[3]
|
505
|
-
pxor @t[7], @x[5]
|
506
|
-
pxor @x[3], @t[6]
|
507
|
-
movdqa @t[0], @x[3]
|
508
|
-
pxor @t[2], @x[6]
|
509
|
-
movdqa @t[6], @x[2]
|
510
|
-
movdqa @t[1], @x[7]
|
511
|
-
movdqa @x[6], @x[4]
|
512
|
-
movdqa @t[3], @x[6]
|
513
|
-
___
|
514
|
-
}
|
515
|
-
|
516
|
-
sub InvMixColumns_orig {
|
517
|
-
my @x=@_[0..7];
|
518
|
-
my @t=@_[8..15];
|
519
|
-
|
520
|
-
$code.=<<___;
|
521
|
-
# multiplication by 0x0e
|
522
|
-
pshufd \$0x93, @x[7], @t[7]
|
523
|
-
movdqa @x[2], @t[2]
|
524
|
-
pxor @x[5], @x[7] # 7 5
|
525
|
-
pxor @x[5], @x[2] # 2 5
|
526
|
-
pshufd \$0x93, @x[0], @t[0]
|
527
|
-
movdqa @x[5], @t[5]
|
528
|
-
pxor @x[0], @x[5] # 5 0 [1]
|
529
|
-
pxor @x[1], @x[0] # 0 1
|
530
|
-
pshufd \$0x93, @x[1], @t[1]
|
531
|
-
pxor @x[2], @x[1] # 1 25
|
532
|
-
pxor @x[6], @x[0] # 01 6 [2]
|
533
|
-
pxor @x[3], @x[1] # 125 3 [4]
|
534
|
-
pshufd \$0x93, @x[3], @t[3]
|
535
|
-
pxor @x[0], @x[2] # 25 016 [3]
|
536
|
-
pxor @x[7], @x[3] # 3 75
|
537
|
-
pxor @x[6], @x[7] # 75 6 [0]
|
538
|
-
pshufd \$0x93, @x[6], @t[6]
|
539
|
-
movdqa @x[4], @t[4]
|
540
|
-
pxor @x[4], @x[6] # 6 4
|
541
|
-
pxor @x[3], @x[4] # 4 375 [6]
|
542
|
-
pxor @x[7], @x[3] # 375 756=36
|
543
|
-
pxor @t[5], @x[6] # 64 5 [7]
|
544
|
-
pxor @t[2], @x[3] # 36 2
|
545
|
-
pxor @t[4], @x[3] # 362 4 [5]
|
546
|
-
pshufd \$0x93, @t[5], @t[5]
|
547
|
-
___
|
548
|
-
my @y = @x[7,5,0,2,1,3,4,6];
|
549
|
-
$code.=<<___;
|
550
|
-
# multiplication by 0x0b
|
551
|
-
pxor @y[0], @y[1]
|
552
|
-
pxor @t[0], @y[0]
|
553
|
-
pxor @t[1], @y[1]
|
554
|
-
pshufd \$0x93, @t[2], @t[2]
|
555
|
-
pxor @t[5], @y[0]
|
556
|
-
pxor @t[6], @y[1]
|
557
|
-
pxor @t[7], @y[0]
|
558
|
-
pshufd \$0x93, @t[4], @t[4]
|
559
|
-
pxor @t[6], @t[7] # clobber t[7]
|
560
|
-
pxor @y[0], @y[1]
|
561
|
-
|
562
|
-
pxor @t[0], @y[3]
|
563
|
-
pshufd \$0x93, @t[0], @t[0]
|
564
|
-
pxor @t[1], @y[2]
|
565
|
-
pxor @t[1], @y[4]
|
566
|
-
pxor @t[2], @y[2]
|
567
|
-
pshufd \$0x93, @t[1], @t[1]
|
568
|
-
pxor @t[2], @y[3]
|
569
|
-
pxor @t[2], @y[5]
|
570
|
-
pxor @t[7], @y[2]
|
571
|
-
pshufd \$0x93, @t[2], @t[2]
|
572
|
-
pxor @t[3], @y[3]
|
573
|
-
pxor @t[3], @y[6]
|
574
|
-
pxor @t[3], @y[4]
|
575
|
-
pshufd \$0x93, @t[3], @t[3]
|
576
|
-
pxor @t[4], @y[7]
|
577
|
-
pxor @t[4], @y[5]
|
578
|
-
pxor @t[7], @y[7]
|
579
|
-
pxor @t[5], @y[3]
|
580
|
-
pxor @t[4], @y[4]
|
581
|
-
pxor @t[5], @t[7] # clobber t[7] even more
|
582
|
-
|
583
|
-
pxor @t[7], @y[5]
|
584
|
-
pshufd \$0x93, @t[4], @t[4]
|
585
|
-
pxor @t[7], @y[6]
|
586
|
-
pxor @t[7], @y[4]
|
587
|
-
|
588
|
-
pxor @t[5], @t[7]
|
589
|
-
pshufd \$0x93, @t[5], @t[5]
|
590
|
-
pxor @t[6], @t[7] # restore t[7]
|
591
|
-
|
592
|
-
# multiplication by 0x0d
|
593
|
-
pxor @y[7], @y[4]
|
594
|
-
pxor @t[4], @y[7]
|
595
|
-
pshufd \$0x93, @t[6], @t[6]
|
596
|
-
pxor @t[0], @y[2]
|
597
|
-
pxor @t[5], @y[7]
|
598
|
-
pxor @t[2], @y[2]
|
599
|
-
pshufd \$0x93, @t[7], @t[7]
|
600
|
-
|
601
|
-
pxor @y[1], @y[3]
|
602
|
-
pxor @t[1], @y[1]
|
603
|
-
pxor @t[0], @y[0]
|
604
|
-
pxor @t[0], @y[3]
|
605
|
-
pxor @t[5], @y[1]
|
606
|
-
pxor @t[5], @y[0]
|
607
|
-
pxor @t[7], @y[1]
|
608
|
-
pshufd \$0x93, @t[0], @t[0]
|
609
|
-
pxor @t[6], @y[0]
|
610
|
-
pxor @y[1], @y[3]
|
611
|
-
pxor @t[1], @y[4]
|
612
|
-
pshufd \$0x93, @t[1], @t[1]
|
613
|
-
|
614
|
-
pxor @t[7], @y[7]
|
615
|
-
pxor @t[2], @y[4]
|
616
|
-
pxor @t[2], @y[5]
|
617
|
-
pshufd \$0x93, @t[2], @t[2]
|
618
|
-
pxor @t[6], @y[2]
|
619
|
-
pxor @t[3], @t[6] # clobber t[6]
|
620
|
-
pxor @y[7], @y[4]
|
621
|
-
pxor @t[6], @y[3]
|
622
|
-
|
623
|
-
pxor @t[6], @y[6]
|
624
|
-
pxor @t[5], @y[5]
|
625
|
-
pxor @t[4], @y[6]
|
626
|
-
pshufd \$0x93, @t[4], @t[4]
|
627
|
-
pxor @t[6], @y[5]
|
628
|
-
pxor @t[7], @y[6]
|
629
|
-
pxor @t[3], @t[6] # restore t[6]
|
630
|
-
|
631
|
-
pshufd \$0x93, @t[5], @t[5]
|
632
|
-
pshufd \$0x93, @t[6], @t[6]
|
633
|
-
pshufd \$0x93, @t[7], @t[7]
|
634
|
-
pshufd \$0x93, @t[3], @t[3]
|
635
|
-
|
636
|
-
# multiplication by 0x09
|
637
|
-
pxor @y[1], @y[4]
|
638
|
-
pxor @y[1], @t[1] # t[1]=y[1]
|
639
|
-
pxor @t[5], @t[0] # clobber t[0]
|
640
|
-
pxor @t[5], @t[1]
|
641
|
-
pxor @t[0], @y[3]
|
642
|
-
pxor @y[0], @t[0] # t[0]=y[0]
|
643
|
-
pxor @t[6], @t[1]
|
644
|
-
pxor @t[7], @t[6] # clobber t[6]
|
645
|
-
pxor @t[1], @y[4]
|
646
|
-
pxor @t[4], @y[7]
|
647
|
-
pxor @y[4], @t[4] # t[4]=y[4]
|
648
|
-
pxor @t[3], @y[6]
|
649
|
-
pxor @y[3], @t[3] # t[3]=y[3]
|
650
|
-
pxor @t[2], @y[5]
|
651
|
-
pxor @y[2], @t[2] # t[2]=y[2]
|
652
|
-
pxor @t[7], @t[3]
|
653
|
-
pxor @y[5], @t[5] # t[5]=y[5]
|
654
|
-
pxor @t[6], @t[2]
|
655
|
-
pxor @t[6], @t[5]
|
656
|
-
pxor @y[6], @t[6] # t[6]=y[6]
|
657
|
-
pxor @y[7], @t[7] # t[7]=y[7]
|
658
|
-
|
659
|
-
movdqa @t[0],@XMM[0]
|
660
|
-
movdqa @t[1],@XMM[1]
|
661
|
-
movdqa @t[2],@XMM[2]
|
662
|
-
movdqa @t[3],@XMM[3]
|
663
|
-
movdqa @t[4],@XMM[4]
|
664
|
-
movdqa @t[5],@XMM[5]
|
665
|
-
movdqa @t[6],@XMM[6]
|
666
|
-
movdqa @t[7],@XMM[7]
|
667
|
-
___
|
668
|
-
}
|
669
|
-
|
670
|
-
sub InvMixColumns {
|
671
|
-
my @x=@_[0..7];
|
672
|
-
my @t=@_[8..15];
|
673
|
-
|
674
|
-
# Thanks to Jussi Kivilinna for providing pointer to
|
675
|
-
#
|
676
|
-
# | 0e 0b 0d 09 | | 02 03 01 01 | | 05 00 04 00 |
|
677
|
-
# | 09 0e 0b 0d | = | 01 02 03 01 | x | 00 05 00 04 |
|
678
|
-
# | 0d 09 0e 0b | | 01 01 02 03 | | 04 00 05 00 |
|
679
|
-
# | 0b 0d 09 0e | | 03 01 01 02 | | 00 04 00 05 |
|
680
|
-
|
681
|
-
$code.=<<___;
|
682
|
-
# multiplication by 0x05-0x00-0x04-0x00
|
683
|
-
pshufd \$0x4E, @x[0], @t[0]
|
684
|
-
pshufd \$0x4E, @x[6], @t[6]
|
685
|
-
pxor @x[0], @t[0]
|
686
|
-
pshufd \$0x4E, @x[7], @t[7]
|
687
|
-
pxor @x[6], @t[6]
|
688
|
-
pshufd \$0x4E, @x[1], @t[1]
|
689
|
-
pxor @x[7], @t[7]
|
690
|
-
pshufd \$0x4E, @x[2], @t[2]
|
691
|
-
pxor @x[1], @t[1]
|
692
|
-
pshufd \$0x4E, @x[3], @t[3]
|
693
|
-
pxor @x[2], @t[2]
|
694
|
-
pxor @t[6], @x[0]
|
695
|
-
pxor @t[6], @x[1]
|
696
|
-
pshufd \$0x4E, @x[4], @t[4]
|
697
|
-
pxor @x[3], @t[3]
|
698
|
-
pxor @t[0], @x[2]
|
699
|
-
pxor @t[1], @x[3]
|
700
|
-
pshufd \$0x4E, @x[5], @t[5]
|
701
|
-
pxor @x[4], @t[4]
|
702
|
-
pxor @t[7], @x[1]
|
703
|
-
pxor @t[2], @x[4]
|
704
|
-
pxor @x[5], @t[5]
|
705
|
-
|
706
|
-
pxor @t[7], @x[2]
|
707
|
-
pxor @t[6], @x[3]
|
708
|
-
pxor @t[6], @x[4]
|
709
|
-
pxor @t[3], @x[5]
|
710
|
-
pxor @t[4], @x[6]
|
711
|
-
pxor @t[7], @x[4]
|
712
|
-
pxor @t[7], @x[5]
|
713
|
-
pxor @t[5], @x[7]
|
714
|
-
___
|
715
|
-
&MixColumns (@x,@t,1); # flipped 2<->3 and 4<->6
|
716
|
-
}
|
717
|
-
|
718
|
-
sub aesenc { # not used
|
719
|
-
my @b=@_[0..7];
|
720
|
-
my @t=@_[8..15];
|
721
|
-
$code.=<<___;
|
722
|
-
movdqa 0x30($const),@t[0] # .LSR
|
723
|
-
___
|
724
|
-
&ShiftRows (@b,@t[0]);
|
725
|
-
&Sbox (@b,@t);
|
726
|
-
&MixColumns (@b[0,1,4,6,3,7,2,5],@t);
|
727
|
-
}
|
728
|
-
|
729
|
-
sub aesenclast { # not used
|
730
|
-
my @b=@_[0..7];
|
731
|
-
my @t=@_[8..15];
|
732
|
-
$code.=<<___;
|
733
|
-
movdqa 0x40($const),@t[0] # .LSRM0
|
734
|
-
___
|
735
|
-
&ShiftRows (@b,@t[0]);
|
736
|
-
&Sbox (@b,@t);
|
737
|
-
$code.=<<___
|
738
|
-
pxor 0x00($key),@b[0]
|
739
|
-
pxor 0x10($key),@b[1]
|
740
|
-
pxor 0x20($key),@b[4]
|
741
|
-
pxor 0x30($key),@b[6]
|
742
|
-
pxor 0x40($key),@b[3]
|
743
|
-
pxor 0x50($key),@b[7]
|
744
|
-
pxor 0x60($key),@b[2]
|
745
|
-
pxor 0x70($key),@b[5]
|
746
|
-
___
|
747
|
-
}
|
748
|
-
|
749
|
-
sub swapmove {
|
750
|
-
my ($a,$b,$n,$mask,$t)=@_;
|
751
|
-
$code.=<<___;
|
752
|
-
movdqa $b,$t
|
753
|
-
psrlq \$$n,$b
|
754
|
-
pxor $a,$b
|
755
|
-
pand $mask,$b
|
756
|
-
pxor $b,$a
|
757
|
-
psllq \$$n,$b
|
758
|
-
pxor $t,$b
|
759
|
-
___
|
760
|
-
}
|
761
|
-
sub swapmove2x {
|
762
|
-
my ($a0,$b0,$a1,$b1,$n,$mask,$t0,$t1)=@_;
|
763
|
-
$code.=<<___;
|
764
|
-
movdqa $b0,$t0
|
765
|
-
psrlq \$$n,$b0
|
766
|
-
movdqa $b1,$t1
|
767
|
-
psrlq \$$n,$b1
|
768
|
-
pxor $a0,$b0
|
769
|
-
pxor $a1,$b1
|
770
|
-
pand $mask,$b0
|
771
|
-
pand $mask,$b1
|
772
|
-
pxor $b0,$a0
|
773
|
-
psllq \$$n,$b0
|
774
|
-
pxor $b1,$a1
|
775
|
-
psllq \$$n,$b1
|
776
|
-
pxor $t0,$b0
|
777
|
-
pxor $t1,$b1
|
778
|
-
___
|
779
|
-
}
|
780
|
-
|
781
|
-
sub bitslice {
|
782
|
-
my @x=reverse(@_[0..7]);
|
783
|
-
my ($t0,$t1,$t2,$t3)=@_[8..11];
|
784
|
-
$code.=<<___;
|
785
|
-
movdqa 0x00($const),$t0 # .LBS0
|
786
|
-
movdqa 0x10($const),$t1 # .LBS1
|
787
|
-
___
|
788
|
-
&swapmove2x(@x[0,1,2,3],1,$t0,$t2,$t3);
|
789
|
-
&swapmove2x(@x[4,5,6,7],1,$t0,$t2,$t3);
|
790
|
-
$code.=<<___;
|
791
|
-
movdqa 0x20($const),$t0 # .LBS2
|
792
|
-
___
|
793
|
-
&swapmove2x(@x[0,2,1,3],2,$t1,$t2,$t3);
|
794
|
-
&swapmove2x(@x[4,6,5,7],2,$t1,$t2,$t3);
|
795
|
-
|
796
|
-
&swapmove2x(@x[0,4,1,5],4,$t0,$t2,$t3);
|
797
|
-
&swapmove2x(@x[2,6,3,7],4,$t0,$t2,$t3);
|
798
|
-
}
|
799
|
-
|
800
|
-
$code.=<<___;
|
801
|
-
.text
|
802
|
-
|
803
|
-
.extern asm_AES_encrypt
|
804
|
-
.extern asm_AES_decrypt
|
805
|
-
|
806
|
-
.type _bsaes_encrypt8,\@abi-omnipotent
|
807
|
-
.align 64
|
808
|
-
_bsaes_encrypt8:
|
809
|
-
lea .LBS0(%rip), $const # constants table
|
810
|
-
|
811
|
-
movdqa ($key), @XMM[9] # round 0 key
|
812
|
-
lea 0x10($key), $key
|
813
|
-
movdqa 0x50($const), @XMM[8] # .LM0SR
|
814
|
-
pxor @XMM[9], @XMM[0] # xor with round0 key
|
815
|
-
pxor @XMM[9], @XMM[1]
|
816
|
-
pxor @XMM[9], @XMM[2]
|
817
|
-
pxor @XMM[9], @XMM[3]
|
818
|
-
pshufb @XMM[8], @XMM[0]
|
819
|
-
pshufb @XMM[8], @XMM[1]
|
820
|
-
pxor @XMM[9], @XMM[4]
|
821
|
-
pxor @XMM[9], @XMM[5]
|
822
|
-
pshufb @XMM[8], @XMM[2]
|
823
|
-
pshufb @XMM[8], @XMM[3]
|
824
|
-
pxor @XMM[9], @XMM[6]
|
825
|
-
pxor @XMM[9], @XMM[7]
|
826
|
-
pshufb @XMM[8], @XMM[4]
|
827
|
-
pshufb @XMM[8], @XMM[5]
|
828
|
-
pshufb @XMM[8], @XMM[6]
|
829
|
-
pshufb @XMM[8], @XMM[7]
|
830
|
-
_bsaes_encrypt8_bitslice:
|
831
|
-
___
|
832
|
-
&bitslice (@XMM[0..7, 8..11]);
|
833
|
-
$code.=<<___;
|
834
|
-
dec $rounds
|
835
|
-
jmp .Lenc_sbox
|
836
|
-
.align 16
|
837
|
-
.Lenc_loop:
|
838
|
-
___
|
839
|
-
&ShiftRows (@XMM[0..7, 8]);
|
840
|
-
$code.=".Lenc_sbox:\n";
|
841
|
-
&Sbox (@XMM[0..7, 8..15]);
|
842
|
-
$code.=<<___;
|
843
|
-
dec $rounds
|
844
|
-
jl .Lenc_done
|
845
|
-
___
|
846
|
-
&MixColumns (@XMM[0,1,4,6,3,7,2,5, 8..15]);
|
847
|
-
$code.=<<___;
|
848
|
-
movdqa 0x30($const), @XMM[8] # .LSR
|
849
|
-
jnz .Lenc_loop
|
850
|
-
movdqa 0x40($const), @XMM[8] # .LSRM0
|
851
|
-
jmp .Lenc_loop
|
852
|
-
.align 16
|
853
|
-
.Lenc_done:
|
854
|
-
___
|
855
|
-
# output in lsb > [t0, t1, t4, t6, t3, t7, t2, t5] < msb
|
856
|
-
&bitslice (@XMM[0,1,4,6,3,7,2,5, 8..11]);
|
857
|
-
$code.=<<___;
|
858
|
-
movdqa ($key), @XMM[8] # last round key
|
859
|
-
pxor @XMM[8], @XMM[4]
|
860
|
-
pxor @XMM[8], @XMM[6]
|
861
|
-
pxor @XMM[8], @XMM[3]
|
862
|
-
pxor @XMM[8], @XMM[7]
|
863
|
-
pxor @XMM[8], @XMM[2]
|
864
|
-
pxor @XMM[8], @XMM[5]
|
865
|
-
pxor @XMM[8], @XMM[0]
|
866
|
-
pxor @XMM[8], @XMM[1]
|
867
|
-
ret
|
868
|
-
.size _bsaes_encrypt8,.-_bsaes_encrypt8
|
869
|
-
|
870
|
-
.type _bsaes_decrypt8,\@abi-omnipotent
|
871
|
-
.align 64
|
872
|
-
_bsaes_decrypt8:
|
873
|
-
lea .LBS0(%rip), $const # constants table
|
874
|
-
|
875
|
-
movdqa ($key), @XMM[9] # round 0 key
|
876
|
-
lea 0x10($key), $key
|
877
|
-
movdqa -0x30($const), @XMM[8] # .LM0ISR
|
878
|
-
pxor @XMM[9], @XMM[0] # xor with round0 key
|
879
|
-
pxor @XMM[9], @XMM[1]
|
880
|
-
pxor @XMM[9], @XMM[2]
|
881
|
-
pxor @XMM[9], @XMM[3]
|
882
|
-
pshufb @XMM[8], @XMM[0]
|
883
|
-
pshufb @XMM[8], @XMM[1]
|
884
|
-
pxor @XMM[9], @XMM[4]
|
885
|
-
pxor @XMM[9], @XMM[5]
|
886
|
-
pshufb @XMM[8], @XMM[2]
|
887
|
-
pshufb @XMM[8], @XMM[3]
|
888
|
-
pxor @XMM[9], @XMM[6]
|
889
|
-
pxor @XMM[9], @XMM[7]
|
890
|
-
pshufb @XMM[8], @XMM[4]
|
891
|
-
pshufb @XMM[8], @XMM[5]
|
892
|
-
pshufb @XMM[8], @XMM[6]
|
893
|
-
pshufb @XMM[8], @XMM[7]
|
894
|
-
___
|
895
|
-
&bitslice (@XMM[0..7, 8..11]);
|
896
|
-
$code.=<<___;
|
897
|
-
dec $rounds
|
898
|
-
jmp .Ldec_sbox
|
899
|
-
.align 16
|
900
|
-
.Ldec_loop:
|
901
|
-
___
|
902
|
-
&ShiftRows (@XMM[0..7, 8]);
|
903
|
-
$code.=".Ldec_sbox:\n";
|
904
|
-
&InvSbox (@XMM[0..7, 8..15]);
|
905
|
-
$code.=<<___;
|
906
|
-
dec $rounds
|
907
|
-
jl .Ldec_done
|
908
|
-
___
|
909
|
-
&InvMixColumns (@XMM[0,1,6,4,2,7,3,5, 8..15]);
|
910
|
-
$code.=<<___;
|
911
|
-
movdqa -0x10($const), @XMM[8] # .LISR
|
912
|
-
jnz .Ldec_loop
|
913
|
-
movdqa -0x20($const), @XMM[8] # .LISRM0
|
914
|
-
jmp .Ldec_loop
|
915
|
-
.align 16
|
916
|
-
.Ldec_done:
|
917
|
-
___
|
918
|
-
&bitslice (@XMM[0,1,6,4,2,7,3,5, 8..11]);
|
919
|
-
$code.=<<___;
|
920
|
-
movdqa ($key), @XMM[8] # last round key
|
921
|
-
pxor @XMM[8], @XMM[6]
|
922
|
-
pxor @XMM[8], @XMM[4]
|
923
|
-
pxor @XMM[8], @XMM[2]
|
924
|
-
pxor @XMM[8], @XMM[7]
|
925
|
-
pxor @XMM[8], @XMM[3]
|
926
|
-
pxor @XMM[8], @XMM[5]
|
927
|
-
pxor @XMM[8], @XMM[0]
|
928
|
-
pxor @XMM[8], @XMM[1]
|
929
|
-
ret
|
930
|
-
.size _bsaes_decrypt8,.-_bsaes_decrypt8
|
931
|
-
___
|
932
|
-
}
|
933
|
-
{
|
934
|
-
my ($out,$inp,$rounds,$const)=("%rax","%rcx","%r10d","%r11");
|
935
|
-
|
936
|
-
sub bitslice_key {
|
937
|
-
my @x=reverse(@_[0..7]);
|
938
|
-
my ($bs0,$bs1,$bs2,$t2,$t3)=@_[8..12];
|
939
|
-
|
940
|
-
&swapmove (@x[0,1],1,$bs0,$t2,$t3);
|
941
|
-
$code.=<<___;
|
942
|
-
#&swapmove(@x[2,3],1,$t0,$t2,$t3);
|
943
|
-
movdqa @x[0], @x[2]
|
944
|
-
movdqa @x[1], @x[3]
|
945
|
-
___
|
946
|
-
#&swapmove2x(@x[4,5,6,7],1,$t0,$t2,$t3);
|
947
|
-
|
948
|
-
&swapmove2x (@x[0,2,1,3],2,$bs1,$t2,$t3);
|
949
|
-
$code.=<<___;
|
950
|
-
#&swapmove2x(@x[4,6,5,7],2,$t1,$t2,$t3);
|
951
|
-
movdqa @x[0], @x[4]
|
952
|
-
movdqa @x[2], @x[6]
|
953
|
-
movdqa @x[1], @x[5]
|
954
|
-
movdqa @x[3], @x[7]
|
955
|
-
___
|
956
|
-
&swapmove2x (@x[0,4,1,5],4,$bs2,$t2,$t3);
|
957
|
-
&swapmove2x (@x[2,6,3,7],4,$bs2,$t2,$t3);
|
958
|
-
}
|
959
|
-
|
960
|
-
$code.=<<___;
|
961
|
-
.type _bsaes_key_convert,\@abi-omnipotent
|
962
|
-
.align 16
|
963
|
-
_bsaes_key_convert:
|
964
|
-
lea .Lmasks(%rip), $const
|
965
|
-
movdqu ($inp), %xmm7 # load round 0 key
|
966
|
-
lea 0x10($inp), $inp
|
967
|
-
movdqa 0x00($const), %xmm0 # 0x01...
|
968
|
-
movdqa 0x10($const), %xmm1 # 0x02...
|
969
|
-
movdqa 0x20($const), %xmm2 # 0x04...
|
970
|
-
movdqa 0x30($const), %xmm3 # 0x08...
|
971
|
-
movdqa 0x40($const), %xmm4 # .LM0
|
972
|
-
pcmpeqd %xmm5, %xmm5 # .LNOT
|
973
|
-
|
974
|
-
movdqu ($inp), %xmm6 # load round 1 key
|
975
|
-
movdqa %xmm7, ($out) # save round 0 key
|
976
|
-
lea 0x10($out), $out
|
977
|
-
dec $rounds
|
978
|
-
jmp .Lkey_loop
|
979
|
-
.align 16
|
980
|
-
.Lkey_loop:
|
981
|
-
pshufb %xmm4, %xmm6 # .LM0
|
982
|
-
|
983
|
-
movdqa %xmm0, %xmm8
|
984
|
-
movdqa %xmm1, %xmm9
|
985
|
-
|
986
|
-
pand %xmm6, %xmm8
|
987
|
-
pand %xmm6, %xmm9
|
988
|
-
movdqa %xmm2, %xmm10
|
989
|
-
pcmpeqb %xmm0, %xmm8
|
990
|
-
psllq \$4, %xmm0 # 0x10...
|
991
|
-
movdqa %xmm3, %xmm11
|
992
|
-
pcmpeqb %xmm1, %xmm9
|
993
|
-
psllq \$4, %xmm1 # 0x20...
|
994
|
-
|
995
|
-
pand %xmm6, %xmm10
|
996
|
-
pand %xmm6, %xmm11
|
997
|
-
movdqa %xmm0, %xmm12
|
998
|
-
pcmpeqb %xmm2, %xmm10
|
999
|
-
psllq \$4, %xmm2 # 0x40...
|
1000
|
-
movdqa %xmm1, %xmm13
|
1001
|
-
pcmpeqb %xmm3, %xmm11
|
1002
|
-
psllq \$4, %xmm3 # 0x80...
|
1003
|
-
|
1004
|
-
movdqa %xmm2, %xmm14
|
1005
|
-
movdqa %xmm3, %xmm15
|
1006
|
-
pxor %xmm5, %xmm8 # "pnot"
|
1007
|
-
pxor %xmm5, %xmm9
|
1008
|
-
|
1009
|
-
pand %xmm6, %xmm12
|
1010
|
-
pand %xmm6, %xmm13
|
1011
|
-
movdqa %xmm8, 0x00($out) # write bit-sliced round key
|
1012
|
-
pcmpeqb %xmm0, %xmm12
|
1013
|
-
psrlq \$4, %xmm0 # 0x01...
|
1014
|
-
movdqa %xmm9, 0x10($out)
|
1015
|
-
pcmpeqb %xmm1, %xmm13
|
1016
|
-
psrlq \$4, %xmm1 # 0x02...
|
1017
|
-
lea 0x10($inp), $inp
|
1018
|
-
|
1019
|
-
pand %xmm6, %xmm14
|
1020
|
-
pand %xmm6, %xmm15
|
1021
|
-
movdqa %xmm10, 0x20($out)
|
1022
|
-
pcmpeqb %xmm2, %xmm14
|
1023
|
-
psrlq \$4, %xmm2 # 0x04...
|
1024
|
-
movdqa %xmm11, 0x30($out)
|
1025
|
-
pcmpeqb %xmm3, %xmm15
|
1026
|
-
psrlq \$4, %xmm3 # 0x08...
|
1027
|
-
movdqu ($inp), %xmm6 # load next round key
|
1028
|
-
|
1029
|
-
pxor %xmm5, %xmm13 # "pnot"
|
1030
|
-
pxor %xmm5, %xmm14
|
1031
|
-
movdqa %xmm12, 0x40($out)
|
1032
|
-
movdqa %xmm13, 0x50($out)
|
1033
|
-
movdqa %xmm14, 0x60($out)
|
1034
|
-
movdqa %xmm15, 0x70($out)
|
1035
|
-
lea 0x80($out),$out
|
1036
|
-
dec $rounds
|
1037
|
-
jnz .Lkey_loop
|
1038
|
-
|
1039
|
-
movdqa 0x50($const), %xmm7 # .L63
|
1040
|
-
#movdqa %xmm6, ($out) # don't save last round key
|
1041
|
-
ret
|
1042
|
-
.size _bsaes_key_convert,.-_bsaes_key_convert
|
1043
|
-
___
|
1044
|
-
}
|
1045
|
-
|
1046
|
-
if (0 && !$win64) { # following four functions are unsupported interface
|
1047
|
-
# used for benchmarking...
|
1048
|
-
$code.=<<___;
|
1049
|
-
.globl bsaes_enc_key_convert
|
1050
|
-
.type bsaes_enc_key_convert,\@function,2
|
1051
|
-
.align 16
|
1052
|
-
bsaes_enc_key_convert:
|
1053
|
-
mov 240($inp),%r10d # pass rounds
|
1054
|
-
mov $inp,%rcx # pass key
|
1055
|
-
mov $out,%rax # pass key schedule
|
1056
|
-
call _bsaes_key_convert
|
1057
|
-
pxor %xmm6,%xmm7 # fix up last round key
|
1058
|
-
movdqa %xmm7,(%rax) # save last round key
|
1059
|
-
ret
|
1060
|
-
.size bsaes_enc_key_convert,.-bsaes_enc_key_convert
|
1061
|
-
|
1062
|
-
.globl bsaes_encrypt_128
|
1063
|
-
.type bsaes_encrypt_128,\@function,4
|
1064
|
-
.align 16
|
1065
|
-
bsaes_encrypt_128:
|
1066
|
-
.Lenc128_loop:
|
1067
|
-
movdqu 0x00($inp), @XMM[0] # load input
|
1068
|
-
movdqu 0x10($inp), @XMM[1]
|
1069
|
-
movdqu 0x20($inp), @XMM[2]
|
1070
|
-
movdqu 0x30($inp), @XMM[3]
|
1071
|
-
movdqu 0x40($inp), @XMM[4]
|
1072
|
-
movdqu 0x50($inp), @XMM[5]
|
1073
|
-
movdqu 0x60($inp), @XMM[6]
|
1074
|
-
movdqu 0x70($inp), @XMM[7]
|
1075
|
-
mov $key, %rax # pass the $key
|
1076
|
-
lea 0x80($inp), $inp
|
1077
|
-
mov \$10,%r10d
|
1078
|
-
|
1079
|
-
call _bsaes_encrypt8
|
1080
|
-
|
1081
|
-
movdqu @XMM[0], 0x00($out) # write output
|
1082
|
-
movdqu @XMM[1], 0x10($out)
|
1083
|
-
movdqu @XMM[4], 0x20($out)
|
1084
|
-
movdqu @XMM[6], 0x30($out)
|
1085
|
-
movdqu @XMM[3], 0x40($out)
|
1086
|
-
movdqu @XMM[7], 0x50($out)
|
1087
|
-
movdqu @XMM[2], 0x60($out)
|
1088
|
-
movdqu @XMM[5], 0x70($out)
|
1089
|
-
lea 0x80($out), $out
|
1090
|
-
sub \$0x80,$len
|
1091
|
-
ja .Lenc128_loop
|
1092
|
-
ret
|
1093
|
-
.size bsaes_encrypt_128,.-bsaes_encrypt_128
|
1094
|
-
|
1095
|
-
.globl bsaes_dec_key_convert
|
1096
|
-
.type bsaes_dec_key_convert,\@function,2
|
1097
|
-
.align 16
|
1098
|
-
bsaes_dec_key_convert:
|
1099
|
-
mov 240($inp),%r10d # pass rounds
|
1100
|
-
mov $inp,%rcx # pass key
|
1101
|
-
mov $out,%rax # pass key schedule
|
1102
|
-
call _bsaes_key_convert
|
1103
|
-
pxor ($out),%xmm7 # fix up round 0 key
|
1104
|
-
movdqa %xmm6,(%rax) # save last round key
|
1105
|
-
movdqa %xmm7,($out)
|
1106
|
-
ret
|
1107
|
-
.size bsaes_dec_key_convert,.-bsaes_dec_key_convert
|
1108
|
-
|
1109
|
-
.globl bsaes_decrypt_128
|
1110
|
-
.type bsaes_decrypt_128,\@function,4
|
1111
|
-
.align 16
|
1112
|
-
bsaes_decrypt_128:
|
1113
|
-
.Ldec128_loop:
|
1114
|
-
movdqu 0x00($inp), @XMM[0] # load input
|
1115
|
-
movdqu 0x10($inp), @XMM[1]
|
1116
|
-
movdqu 0x20($inp), @XMM[2]
|
1117
|
-
movdqu 0x30($inp), @XMM[3]
|
1118
|
-
movdqu 0x40($inp), @XMM[4]
|
1119
|
-
movdqu 0x50($inp), @XMM[5]
|
1120
|
-
movdqu 0x60($inp), @XMM[6]
|
1121
|
-
movdqu 0x70($inp), @XMM[7]
|
1122
|
-
mov $key, %rax # pass the $key
|
1123
|
-
lea 0x80($inp), $inp
|
1124
|
-
mov \$10,%r10d
|
1125
|
-
|
1126
|
-
call _bsaes_decrypt8
|
1127
|
-
|
1128
|
-
movdqu @XMM[0], 0x00($out) # write output
|
1129
|
-
movdqu @XMM[1], 0x10($out)
|
1130
|
-
movdqu @XMM[6], 0x20($out)
|
1131
|
-
movdqu @XMM[4], 0x30($out)
|
1132
|
-
movdqu @XMM[2], 0x40($out)
|
1133
|
-
movdqu @XMM[7], 0x50($out)
|
1134
|
-
movdqu @XMM[3], 0x60($out)
|
1135
|
-
movdqu @XMM[5], 0x70($out)
|
1136
|
-
lea 0x80($out), $out
|
1137
|
-
sub \$0x80,$len
|
1138
|
-
ja .Ldec128_loop
|
1139
|
-
ret
|
1140
|
-
.size bsaes_decrypt_128,.-bsaes_decrypt_128
|
1141
|
-
___
|
1142
|
-
}
|
1143
|
-
{
|
1144
|
-
######################################################################
|
1145
|
-
#
|
1146
|
-
# OpenSSL interface
|
1147
|
-
#
|
1148
|
-
my ($arg1,$arg2,$arg3,$arg4,$arg5,$arg6)=$win64 ? ("%rcx","%rdx","%r8","%r9","%r10","%r11d")
|
1149
|
-
: ("%rdi","%rsi","%rdx","%rcx","%r8","%r9d");
|
1150
|
-
my ($inp,$out,$len,$key)=("%r12","%r13","%r14","%r15");
|
1151
|
-
|
1152
|
-
$code.=<<___;
|
1153
|
-
.globl bsaes_ctr32_encrypt_blocks
|
1154
|
-
.type bsaes_ctr32_encrypt_blocks,\@abi-omnipotent
|
1155
|
-
.align 16
|
1156
|
-
bsaes_ctr32_encrypt_blocks:
|
1157
|
-
mov %rsp, %rax
|
1158
|
-
.Lctr_enc_prologue:
|
1159
|
-
push %rbp
|
1160
|
-
push %rbx
|
1161
|
-
push %r12
|
1162
|
-
push %r13
|
1163
|
-
push %r14
|
1164
|
-
push %r15
|
1165
|
-
lea -0x48(%rsp), %rsp
|
1166
|
-
___
|
1167
|
-
$code.=<<___ if ($win64);
|
1168
|
-
mov 0xa0(%rsp),$arg5 # pull ivp
|
1169
|
-
lea -0xa0(%rsp), %rsp
|
1170
|
-
movaps %xmm6, 0x40(%rsp)
|
1171
|
-
movaps %xmm7, 0x50(%rsp)
|
1172
|
-
movaps %xmm8, 0x60(%rsp)
|
1173
|
-
movaps %xmm9, 0x70(%rsp)
|
1174
|
-
movaps %xmm10, 0x80(%rsp)
|
1175
|
-
movaps %xmm11, 0x90(%rsp)
|
1176
|
-
movaps %xmm12, 0xa0(%rsp)
|
1177
|
-
movaps %xmm13, 0xb0(%rsp)
|
1178
|
-
movaps %xmm14, 0xc0(%rsp)
|
1179
|
-
movaps %xmm15, 0xd0(%rsp)
|
1180
|
-
.Lctr_enc_body:
|
1181
|
-
___
|
1182
|
-
$code.=<<___;
|
1183
|
-
mov %rsp, %rbp # backup %rsp
|
1184
|
-
movdqu ($arg5), %xmm0 # load counter
|
1185
|
-
mov 240($arg4), %eax # rounds
|
1186
|
-
mov $arg1, $inp # backup arguments
|
1187
|
-
mov $arg2, $out
|
1188
|
-
mov $arg3, $len
|
1189
|
-
mov $arg4, $key
|
1190
|
-
movdqa %xmm0, 0x20(%rbp) # copy counter
|
1191
|
-
cmp \$8, $arg3
|
1192
|
-
jb .Lctr_enc_short
|
1193
|
-
|
1194
|
-
mov %eax, %ebx # rounds
|
1195
|
-
shl \$7, %rax # 128 bytes per inner round key
|
1196
|
-
sub \$`128-32`, %rax # size of bit-sliced key schedule
|
1197
|
-
sub %rax, %rsp
|
1198
|
-
|
1199
|
-
mov %rsp, %rax # pass key schedule
|
1200
|
-
mov $key, %rcx # pass key
|
1201
|
-
mov %ebx, %r10d # pass rounds
|
1202
|
-
call _bsaes_key_convert
|
1203
|
-
pxor %xmm6,%xmm7 # fix up last round key
|
1204
|
-
movdqa %xmm7,(%rax) # save last round key
|
1205
|
-
|
1206
|
-
movdqa (%rsp), @XMM[9] # load round0 key
|
1207
|
-
lea .LADD1(%rip), %r11
|
1208
|
-
movdqa 0x20(%rbp), @XMM[0] # counter copy
|
1209
|
-
movdqa -0x20(%r11), @XMM[8] # .LSWPUP
|
1210
|
-
pshufb @XMM[8], @XMM[9] # byte swap upper part
|
1211
|
-
pshufb @XMM[8], @XMM[0]
|
1212
|
-
movdqa @XMM[9], (%rsp) # save adjusted round0 key
|
1213
|
-
jmp .Lctr_enc_loop
|
1214
|
-
.align 16
|
1215
|
-
.Lctr_enc_loop:
|
1216
|
-
movdqa @XMM[0], 0x20(%rbp) # save counter
|
1217
|
-
movdqa @XMM[0], @XMM[1] # prepare 8 counter values
|
1218
|
-
movdqa @XMM[0], @XMM[2]
|
1219
|
-
paddd 0x00(%r11), @XMM[1] # .LADD1
|
1220
|
-
movdqa @XMM[0], @XMM[3]
|
1221
|
-
paddd 0x10(%r11), @XMM[2] # .LADD2
|
1222
|
-
movdqa @XMM[0], @XMM[4]
|
1223
|
-
paddd 0x20(%r11), @XMM[3] # .LADD3
|
1224
|
-
movdqa @XMM[0], @XMM[5]
|
1225
|
-
paddd 0x30(%r11), @XMM[4] # .LADD4
|
1226
|
-
movdqa @XMM[0], @XMM[6]
|
1227
|
-
paddd 0x40(%r11), @XMM[5] # .LADD5
|
1228
|
-
movdqa @XMM[0], @XMM[7]
|
1229
|
-
paddd 0x50(%r11), @XMM[6] # .LADD6
|
1230
|
-
paddd 0x60(%r11), @XMM[7] # .LADD7
|
1231
|
-
|
1232
|
-
# Borrow prologue from _bsaes_encrypt8 to use the opportunity
|
1233
|
-
# to flip byte order in 32-bit counter
|
1234
|
-
movdqa (%rsp), @XMM[9] # round 0 key
|
1235
|
-
lea 0x10(%rsp), %rax # pass key schedule
|
1236
|
-
movdqa -0x10(%r11), @XMM[8] # .LSWPUPM0SR
|
1237
|
-
pxor @XMM[9], @XMM[0] # xor with round0 key
|
1238
|
-
pxor @XMM[9], @XMM[1]
|
1239
|
-
pxor @XMM[9], @XMM[2]
|
1240
|
-
pxor @XMM[9], @XMM[3]
|
1241
|
-
pshufb @XMM[8], @XMM[0]
|
1242
|
-
pshufb @XMM[8], @XMM[1]
|
1243
|
-
pxor @XMM[9], @XMM[4]
|
1244
|
-
pxor @XMM[9], @XMM[5]
|
1245
|
-
pshufb @XMM[8], @XMM[2]
|
1246
|
-
pshufb @XMM[8], @XMM[3]
|
1247
|
-
pxor @XMM[9], @XMM[6]
|
1248
|
-
pxor @XMM[9], @XMM[7]
|
1249
|
-
pshufb @XMM[8], @XMM[4]
|
1250
|
-
pshufb @XMM[8], @XMM[5]
|
1251
|
-
pshufb @XMM[8], @XMM[6]
|
1252
|
-
pshufb @XMM[8], @XMM[7]
|
1253
|
-
lea .LBS0(%rip), %r11 # constants table
|
1254
|
-
mov %ebx,%r10d # pass rounds
|
1255
|
-
|
1256
|
-
call _bsaes_encrypt8_bitslice
|
1257
|
-
|
1258
|
-
sub \$8,$len
|
1259
|
-
jc .Lctr_enc_loop_done
|
1260
|
-
|
1261
|
-
movdqu 0x00($inp), @XMM[8] # load input
|
1262
|
-
movdqu 0x10($inp), @XMM[9]
|
1263
|
-
movdqu 0x20($inp), @XMM[10]
|
1264
|
-
movdqu 0x30($inp), @XMM[11]
|
1265
|
-
movdqu 0x40($inp), @XMM[12]
|
1266
|
-
movdqu 0x50($inp), @XMM[13]
|
1267
|
-
movdqu 0x60($inp), @XMM[14]
|
1268
|
-
movdqu 0x70($inp), @XMM[15]
|
1269
|
-
lea 0x80($inp),$inp
|
1270
|
-
pxor @XMM[0], @XMM[8]
|
1271
|
-
movdqa 0x20(%rbp), @XMM[0] # load counter
|
1272
|
-
pxor @XMM[9], @XMM[1]
|
1273
|
-
movdqu @XMM[8], 0x00($out) # write output
|
1274
|
-
pxor @XMM[10], @XMM[4]
|
1275
|
-
movdqu @XMM[1], 0x10($out)
|
1276
|
-
pxor @XMM[11], @XMM[6]
|
1277
|
-
movdqu @XMM[4], 0x20($out)
|
1278
|
-
pxor @XMM[12], @XMM[3]
|
1279
|
-
movdqu @XMM[6], 0x30($out)
|
1280
|
-
pxor @XMM[13], @XMM[7]
|
1281
|
-
movdqu @XMM[3], 0x40($out)
|
1282
|
-
pxor @XMM[14], @XMM[2]
|
1283
|
-
movdqu @XMM[7], 0x50($out)
|
1284
|
-
pxor @XMM[15], @XMM[5]
|
1285
|
-
movdqu @XMM[2], 0x60($out)
|
1286
|
-
lea .LADD1(%rip), %r11
|
1287
|
-
movdqu @XMM[5], 0x70($out)
|
1288
|
-
lea 0x80($out), $out
|
1289
|
-
paddd 0x70(%r11), @XMM[0] # .LADD8
|
1290
|
-
jnz .Lctr_enc_loop
|
1291
|
-
|
1292
|
-
jmp .Lctr_enc_done
|
1293
|
-
.align 16
|
1294
|
-
.Lctr_enc_loop_done:
|
1295
|
-
add \$8, $len
|
1296
|
-
movdqu 0x00($inp), @XMM[8] # load input
|
1297
|
-
pxor @XMM[8], @XMM[0]
|
1298
|
-
movdqu @XMM[0], 0x00($out) # write output
|
1299
|
-
cmp \$2,$len
|
1300
|
-
jb .Lctr_enc_done
|
1301
|
-
movdqu 0x10($inp), @XMM[9]
|
1302
|
-
pxor @XMM[9], @XMM[1]
|
1303
|
-
movdqu @XMM[1], 0x10($out)
|
1304
|
-
je .Lctr_enc_done
|
1305
|
-
movdqu 0x20($inp), @XMM[10]
|
1306
|
-
pxor @XMM[10], @XMM[4]
|
1307
|
-
movdqu @XMM[4], 0x20($out)
|
1308
|
-
cmp \$4,$len
|
1309
|
-
jb .Lctr_enc_done
|
1310
|
-
movdqu 0x30($inp), @XMM[11]
|
1311
|
-
pxor @XMM[11], @XMM[6]
|
1312
|
-
movdqu @XMM[6], 0x30($out)
|
1313
|
-
je .Lctr_enc_done
|
1314
|
-
movdqu 0x40($inp), @XMM[12]
|
1315
|
-
pxor @XMM[12], @XMM[3]
|
1316
|
-
movdqu @XMM[3], 0x40($out)
|
1317
|
-
cmp \$6,$len
|
1318
|
-
jb .Lctr_enc_done
|
1319
|
-
movdqu 0x50($inp), @XMM[13]
|
1320
|
-
pxor @XMM[13], @XMM[7]
|
1321
|
-
movdqu @XMM[7], 0x50($out)
|
1322
|
-
je .Lctr_enc_done
|
1323
|
-
movdqu 0x60($inp), @XMM[14]
|
1324
|
-
pxor @XMM[14], @XMM[2]
|
1325
|
-
movdqu @XMM[2], 0x60($out)
|
1326
|
-
jmp .Lctr_enc_done
|
1327
|
-
|
1328
|
-
.align 16
|
1329
|
-
.Lctr_enc_short:
|
1330
|
-
lea 0x20(%rbp), $arg1
|
1331
|
-
lea 0x30(%rbp), $arg2
|
1332
|
-
lea ($key), $arg3
|
1333
|
-
call asm_AES_encrypt
|
1334
|
-
movdqu ($inp), @XMM[1]
|
1335
|
-
lea 16($inp), $inp
|
1336
|
-
mov 0x2c(%rbp), %eax # load 32-bit counter
|
1337
|
-
bswap %eax
|
1338
|
-
pxor 0x30(%rbp), @XMM[1]
|
1339
|
-
inc %eax # increment
|
1340
|
-
movdqu @XMM[1], ($out)
|
1341
|
-
bswap %eax
|
1342
|
-
lea 16($out), $out
|
1343
|
-
mov %eax, 0x2c(%rsp) # save 32-bit counter
|
1344
|
-
dec $len
|
1345
|
-
jnz .Lctr_enc_short
|
1346
|
-
|
1347
|
-
.Lctr_enc_done:
|
1348
|
-
lea (%rsp), %rax
|
1349
|
-
pxor %xmm0, %xmm0
|
1350
|
-
.Lctr_enc_bzero: # wipe key schedule [if any]
|
1351
|
-
movdqa %xmm0, 0x00(%rax)
|
1352
|
-
movdqa %xmm0, 0x10(%rax)
|
1353
|
-
lea 0x20(%rax), %rax
|
1354
|
-
cmp %rax, %rbp
|
1355
|
-
ja .Lctr_enc_bzero
|
1356
|
-
|
1357
|
-
lea (%rbp),%rsp # restore %rsp
|
1358
|
-
___
|
1359
|
-
$code.=<<___ if ($win64);
|
1360
|
-
movaps 0x40(%rbp), %xmm6
|
1361
|
-
movaps 0x50(%rbp), %xmm7
|
1362
|
-
movaps 0x60(%rbp), %xmm8
|
1363
|
-
movaps 0x70(%rbp), %xmm9
|
1364
|
-
movaps 0x80(%rbp), %xmm10
|
1365
|
-
movaps 0x90(%rbp), %xmm11
|
1366
|
-
movaps 0xa0(%rbp), %xmm12
|
1367
|
-
movaps 0xb0(%rbp), %xmm13
|
1368
|
-
movaps 0xc0(%rbp), %xmm14
|
1369
|
-
movaps 0xd0(%rbp), %xmm15
|
1370
|
-
lea 0xa0(%rbp), %rsp
|
1371
|
-
___
|
1372
|
-
$code.=<<___;
|
1373
|
-
mov 0x48(%rsp), %r15
|
1374
|
-
mov 0x50(%rsp), %r14
|
1375
|
-
mov 0x58(%rsp), %r13
|
1376
|
-
mov 0x60(%rsp), %r12
|
1377
|
-
mov 0x68(%rsp), %rbx
|
1378
|
-
mov 0x70(%rsp), %rax
|
1379
|
-
lea 0x78(%rsp), %rsp
|
1380
|
-
mov %rax, %rbp
|
1381
|
-
.Lctr_enc_epilogue:
|
1382
|
-
ret
|
1383
|
-
.size bsaes_ctr32_encrypt_blocks,.-bsaes_ctr32_encrypt_blocks
|
1384
|
-
___
|
1385
|
-
}
|
1386
|
-
$code.=<<___;
|
1387
|
-
.type _bsaes_const,\@object
|
1388
|
-
.align 64
|
1389
|
-
_bsaes_const:
|
1390
|
-
.LM0ISR: # InvShiftRows constants
|
1391
|
-
.quad 0x0a0e0206070b0f03, 0x0004080c0d010509
|
1392
|
-
.LISRM0:
|
1393
|
-
.quad 0x01040b0e0205080f, 0x0306090c00070a0d
|
1394
|
-
.LISR:
|
1395
|
-
.quad 0x0504070602010003, 0x0f0e0d0c080b0a09
|
1396
|
-
.LBS0: # bit-slice constants
|
1397
|
-
.quad 0x5555555555555555, 0x5555555555555555
|
1398
|
-
.LBS1:
|
1399
|
-
.quad 0x3333333333333333, 0x3333333333333333
|
1400
|
-
.LBS2:
|
1401
|
-
.quad 0x0f0f0f0f0f0f0f0f, 0x0f0f0f0f0f0f0f0f
|
1402
|
-
.LSR: # shiftrows constants
|
1403
|
-
.quad 0x0504070600030201, 0x0f0e0d0c0a09080b
|
1404
|
-
.LSRM0:
|
1405
|
-
.quad 0x0304090e00050a0f, 0x01060b0c0207080d
|
1406
|
-
.LM0SR:
|
1407
|
-
.quad 0x0a0e02060f03070b, 0x0004080c05090d01
|
1408
|
-
.LSWPUP: # byte-swap upper dword
|
1409
|
-
.quad 0x0706050403020100, 0x0c0d0e0f0b0a0908
|
1410
|
-
.LSWPUPM0SR:
|
1411
|
-
.quad 0x0a0d02060c03070b, 0x0004080f05090e01
|
1412
|
-
.LADD1: # counter increment constants
|
1413
|
-
.quad 0x0000000000000000, 0x0000000100000000
|
1414
|
-
.LADD2:
|
1415
|
-
.quad 0x0000000000000000, 0x0000000200000000
|
1416
|
-
.LADD3:
|
1417
|
-
.quad 0x0000000000000000, 0x0000000300000000
|
1418
|
-
.LADD4:
|
1419
|
-
.quad 0x0000000000000000, 0x0000000400000000
|
1420
|
-
.LADD5:
|
1421
|
-
.quad 0x0000000000000000, 0x0000000500000000
|
1422
|
-
.LADD6:
|
1423
|
-
.quad 0x0000000000000000, 0x0000000600000000
|
1424
|
-
.LADD7:
|
1425
|
-
.quad 0x0000000000000000, 0x0000000700000000
|
1426
|
-
.LADD8:
|
1427
|
-
.quad 0x0000000000000000, 0x0000000800000000
|
1428
|
-
.Lmasks:
|
1429
|
-
.quad 0x0101010101010101, 0x0101010101010101
|
1430
|
-
.quad 0x0202020202020202, 0x0202020202020202
|
1431
|
-
.quad 0x0404040404040404, 0x0404040404040404
|
1432
|
-
.quad 0x0808080808080808, 0x0808080808080808
|
1433
|
-
.LM0:
|
1434
|
-
.quad 0x02060a0e03070b0f, 0x0004080c0105090d
|
1435
|
-
.L63:
|
1436
|
-
.quad 0x6363636363636363, 0x6363636363636363
|
1437
|
-
.asciz "Bit-sliced AES for x86_64/SSSE3, Emilia Käsper, Peter Schwabe, Andy Polyakov"
|
1438
|
-
.align 64
|
1439
|
-
.size _bsaes_const,.-_bsaes_const
|
1440
|
-
___
|
1441
|
-
|
1442
|
-
# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
|
1443
|
-
# CONTEXT *context,DISPATCHER_CONTEXT *disp)
|
1444
|
-
if ($win64) {
|
1445
|
-
$rec="%rcx";
|
1446
|
-
$frame="%rdx";
|
1447
|
-
$context="%r8";
|
1448
|
-
$disp="%r9";
|
1449
|
-
|
1450
|
-
$code.=<<___;
|
1451
|
-
.extern __imp_RtlVirtualUnwind
|
1452
|
-
.type se_handler,\@abi-omnipotent
|
1453
|
-
.align 16
|
1454
|
-
se_handler:
|
1455
|
-
push %rsi
|
1456
|
-
push %rdi
|
1457
|
-
push %rbx
|
1458
|
-
push %rbp
|
1459
|
-
push %r12
|
1460
|
-
push %r13
|
1461
|
-
push %r14
|
1462
|
-
push %r15
|
1463
|
-
pushfq
|
1464
|
-
sub \$64,%rsp
|
1465
|
-
|
1466
|
-
mov 120($context),%rax # pull context->Rax
|
1467
|
-
mov 248($context),%rbx # pull context->Rip
|
1468
|
-
|
1469
|
-
mov 8($disp),%rsi # disp->ImageBase
|
1470
|
-
mov 56($disp),%r11 # disp->HandlerData
|
1471
|
-
|
1472
|
-
mov 0(%r11),%r10d # HandlerData[0]
|
1473
|
-
lea (%rsi,%r10),%r10 # prologue label
|
1474
|
-
cmp %r10,%rbx # context->Rip<prologue label
|
1475
|
-
jb .Lin_prologue
|
1476
|
-
|
1477
|
-
mov 152($context),%rax # pull context->Rsp
|
1478
|
-
|
1479
|
-
mov 4(%r11),%r10d # HandlerData[1]
|
1480
|
-
lea (%rsi,%r10),%r10 # epilogue label
|
1481
|
-
cmp %r10,%rbx # context->Rip>=epilogue label
|
1482
|
-
jae .Lin_prologue
|
1483
|
-
|
1484
|
-
mov 160($context),%rax # pull context->Rbp
|
1485
|
-
|
1486
|
-
lea 0x40(%rax),%rsi # %xmm save area
|
1487
|
-
lea 512($context),%rdi # &context.Xmm6
|
1488
|
-
mov \$20,%ecx # 10*sizeof(%xmm0)/sizeof(%rax)
|
1489
|
-
.long 0xa548f3fc # cld; rep movsq
|
1490
|
-
lea 0xa0(%rax),%rax # adjust stack pointer
|
1491
|
-
|
1492
|
-
mov 0x70(%rax),%rbp
|
1493
|
-
mov 0x68(%rax),%rbx
|
1494
|
-
mov 0x60(%rax),%r12
|
1495
|
-
mov 0x58(%rax),%r13
|
1496
|
-
mov 0x50(%rax),%r14
|
1497
|
-
mov 0x48(%rax),%r15
|
1498
|
-
lea 0x78(%rax),%rax # adjust stack pointer
|
1499
|
-
mov %rbx,144($context) # restore context->Rbx
|
1500
|
-
mov %rbp,160($context) # restore context->Rbp
|
1501
|
-
mov %r12,216($context) # restore context->R12
|
1502
|
-
mov %r13,224($context) # restore context->R13
|
1503
|
-
mov %r14,232($context) # restore context->R14
|
1504
|
-
mov %r15,240($context) # restore context->R15
|
1505
|
-
|
1506
|
-
.Lin_prologue:
|
1507
|
-
mov %rax,152($context) # restore context->Rsp
|
1508
|
-
|
1509
|
-
mov 40($disp),%rdi # disp->ContextRecord
|
1510
|
-
mov $context,%rsi # context
|
1511
|
-
mov \$`1232/8`,%ecx # sizeof(CONTEXT)
|
1512
|
-
.long 0xa548f3fc # cld; rep movsq
|
1513
|
-
|
1514
|
-
mov $disp,%rsi
|
1515
|
-
xor %rcx,%rcx # arg1, UNW_FLAG_NHANDLER
|
1516
|
-
mov 8(%rsi),%rdx # arg2, disp->ImageBase
|
1517
|
-
mov 0(%rsi),%r8 # arg3, disp->ControlPc
|
1518
|
-
mov 16(%rsi),%r9 # arg4, disp->FunctionEntry
|
1519
|
-
mov 40(%rsi),%r10 # disp->ContextRecord
|
1520
|
-
lea 56(%rsi),%r11 # &disp->HandlerData
|
1521
|
-
lea 24(%rsi),%r12 # &disp->EstablisherFrame
|
1522
|
-
mov %r10,32(%rsp) # arg5
|
1523
|
-
mov %r11,40(%rsp) # arg6
|
1524
|
-
mov %r12,48(%rsp) # arg7
|
1525
|
-
mov %rcx,56(%rsp) # arg8, (NULL)
|
1526
|
-
call *__imp_RtlVirtualUnwind(%rip)
|
1527
|
-
|
1528
|
-
mov \$1,%eax # ExceptionContinueSearch
|
1529
|
-
add \$64,%rsp
|
1530
|
-
popfq
|
1531
|
-
pop %r15
|
1532
|
-
pop %r14
|
1533
|
-
pop %r13
|
1534
|
-
pop %r12
|
1535
|
-
pop %rbp
|
1536
|
-
pop %rbx
|
1537
|
-
pop %rdi
|
1538
|
-
pop %rsi
|
1539
|
-
ret
|
1540
|
-
.size se_handler,.-se_handler
|
1541
|
-
|
1542
|
-
.section .pdata
|
1543
|
-
.align 4
|
1544
|
-
___
|
1545
|
-
$code.=<<___;
|
1546
|
-
.rva .Lctr_enc_prologue
|
1547
|
-
.rva .Lctr_enc_epilogue
|
1548
|
-
.rva .Lctr_enc_info
|
1549
|
-
|
1550
|
-
.section .xdata
|
1551
|
-
.align 8
|
1552
|
-
___
|
1553
|
-
$code.=<<___;
|
1554
|
-
.Lctr_enc_info:
|
1555
|
-
.byte 9,0,0,0
|
1556
|
-
.rva se_handler
|
1557
|
-
.rva .Lctr_enc_body,.Lctr_enc_epilogue # HandlerData[]
|
1558
|
-
___
|
1559
|
-
}
|
1560
|
-
|
1561
|
-
$code =~ s/\`([^\`]*)\`/eval($1)/gem;
|
1562
|
-
|
1563
|
-
print $code;
|
1564
|
-
|
1565
|
-
close STDOUT;
|