nettle-bugs March 2018

nettle-bugs@lists.lysator.liu.se

7 participants
9 discussions

[PATCH 0/8] Implement Curve448 ECDH and Ed448
by Daiki Ueno 03 Jan '20

03 Jan '20

Hello, This patch series implements the Curve448 Diffie-Hellman protocol (RFC 7748) and the Ed448 signature scheme (RFC 8032). Although I tried to make it as close as possible to the Curve25519 and Ed25519 implementations, I had to add a few special cases, namely: - for Curve448, eccdata directly calculates points on the Edwards curve instead of the equivalent Montgomery curve - untwisted versions of ecc_add_eh* and ecc_dup_eh are added - the point decoding for Ed448 uses a different formula to recover u and v Also, optimized implementation of modular reduction is currently missing, which is beyond my expertise. I would appreciate any suggestions regarding that. The patches are also available on: https://gitlab.com/dueno/nettle/commits/wip/dueno/ed448 Thanks to Hubert Kario and Nikos Mavrogiannopoulos for initial reviews. Daiki Ueno (8): ecc-mul-test: Fix mpn_cmp calls eccdata: Emit correct ecc_Bmodq_shifted for curve448 eccdata: Redirect ecc_point_out to given stream, instead of stderr ecc: Add add_hh and dup members to ecc_curve ecc-eh-to-a, eddsa-sign: Parameterize hard-coded value Implement curve448 primitives Implement SHAKE128/256 functions Implement Ed448 signature scheme .gitignore | 1 + Makefile.in | 18 +- curve448-eh-to-x.c | 73 + curve448-mul-g.c | 74 + curve448-mul.c | 148 + curve448.h | 58 + ecc-192.c | 5 + ecc-224.c | 5 + ecc-25519.c | 5 + ecc-256.c | 5 + ecc-384.c | 5 + ecc-448.c | 273 ++ ecc-521.c | 5 + ecc-add-eh.c | 74 +- ecc-add-ehh.c | 77 +- ecc-dup-eh.c | 55 +- ecc-eh-to-a.c | 4 +- ecc-internal.h | 57 +- ecc-mul-a-eh.c | 12 +- ecc-mul-g-eh.c | 4 +- ecc-point-mul-g.c | 7 +- ecc-point-mul.c | 2 +- ecc-point.c | 15 + eccdata.c | 183 +- ecdsa-keygen.c | 4 +- ed25519-sha512-sign.c | 15 + ed448-shake256-pubkey.c | 60 + ed448-shake256-sign.c | 92 + ed448-shake256-verify.c | 66 + eddsa-compress.c | 11 +- eddsa-decompress.c | 15 +- eddsa-expand.c | 20 +- eddsa-hash.c | 35 + eddsa-pubkey.c | 2 +- eddsa-sign.c | 18 +- eddsa-verify.c | 16 +- eddsa.h | 24 + examples/ecc-benchmark.c | 1 + nettle-internal.h | 2 +- nettle-meta-hashes.c | 2 + nettle-meta.h | 2 + nettle.texinfo | 152 +- sha3.c | 13 + sha3.h | 56 + shake128-meta.c | 42 + shake128.c | 84 + shake256-meta.c | 42 + shake256.c | 84 + testsuite/.test-rules.make | 12 + testsuite/Makefile.in | 5 +- testsuite/curve448-dh-test.c | 100 + testsuite/ecc-add-test.c | 48 +- testsuite/ecc-dup-test.c | 12 +- testsuite/ecc-mul-a-test.c | 6 +- testsuite/ecc-mul-g-test.c | 6 +- testsuite/ecdh-test.c | 16 +- testsuite/ecdsa-keygen-test.c | 16 + testsuite/ed448-test.c | 240 ++ testsuite/eddsa-compress-test.c | 137 +- testsuite/eddsa-sign-test.c | 66 +- testsuite/eddsa-verify-test.c | 49 +- testsuite/meta-hash-test.c | 2 + testsuite/shake.awk | 14 + testsuite/shake128-test.c | 6183 +++++++++++++++++++++++++++++++++++++++ testsuite/shake256-test.c | 6183 +++++++++++++++++++++++++++++++++++++++ testsuite/testutils.c | 57 +- 66 files changed, 14976 insertions(+), 199 deletions(-) create mode 100644 curve448-eh-to-x.c create mode 100644 curve448-mul-g.c create mode 100644 curve448-mul.c create mode 100644 curve448.h create mode 100644 ecc-448.c create mode 100644 ed448-shake256-pubkey.c create mode 100644 ed448-shake256-sign.c create mode 100644 ed448-shake256-verify.c create mode 100644 shake128-meta.c create mode 100644 shake128.c create mode 100644 shake256-meta.c create mode 100644 shake256.c create mode 100644 testsuite/curve448-dh-test.c create mode 100644 testsuite/ed448-test.c create mode 100755 testsuite/shake.awk create mode 100644 testsuite/shake128-test.c create mode 100644 testsuite/shake256-test.c -- 2.13.3

2 35

API for new AEAD modes
by Nikos Mavrogiannopoulos 31 May '18

31 May '18

Hi, As it is now AEAD ciphers in nettle are supported with their own API. AES-CCM provides: ccm_aes128_set_key ccm_aes128_set_nonce ccm_aes128_update ccm_aes128_encrypt ccm_aes128_decrypt ccm_aes128_digest ccm_aes128_encrypt_message ccm_aes128_decrypt_message AES-GCM: gcm_aes128_set_key gcm_aes128_update gcm_aes128_set_iv gcm_aes128_encrypt gcm_aes128_decrypt gcm_aes128_digest chacha-poly1305: chacha_poly1305_set_key chacha_poly1305_set_nonce chacha_poly1305_update chacha_poly1305_encrypt chacha_poly1305_decrypt chacha_poly1305_digest ccm_aes128_set_nonce takes different parameters from the chacha and gcm equivalent function. Furthermore the AEAD modes for block ciphers include an abstract layer which can be re-used of other ciphers. For example the GCM mode abstraction is re-used for camellia as well. So overall there is a kind of "abstract" API followed with the set_key, set_nonce, update, encrypt/decrypt, however these have not always the same parameters as in the CCM case. I'm checking the AES-SIV-CMAC mode which cannot even be put in that abstraction as the MAC in SIV is the IV and thus the MAC-style API doesn't really suit it. My understanding is that AEAD modes according to RFC5116 need to have 4 inputs: key, nonce, associated data, plaintext so the CCM encrypt and decrypt message seem to be the closer to that API. Would it make sense to standardize on these, and only provide that API for AEAD modes? The reason I'm asking is because SIV could benefit of a very custom API as well because it can take advantage of multiple associated data, but in the end I believe AEAD is about simplicity. Providing a unique API per AEAD cipher seems to me quite contradictory to that goal. regards, Nikos

4 6

symbol versioning update
by Nikos Mavrogiannopoulos 03 Apr '18

03 Apr '18

Hi, When we initially introduced symbol versioning in nettle we bundled all symbols from the library in a single version. That means that new symbols added to a release like nettle_get_hashes() may cause issues like this: https://bugzilla.redhat.com/show_bug.cgi?id=1549190 The underlying issue is in rpm-based systems which detect which symbols exist in which library version based on the version. That way, when a dependency is tracked, the version with the right symbol will be retrieved. Debian-based systems work differently so that shouldn't affect it. To address these problems in gnutls, I follow these additional rules in to having a map file: https://gitlab.com/gnutls/gnutls/blob/master/CONTRIBUTING.md#symbol-and-lib… which in turn comes from libvirt: https://www.berrange.com/posts/2011/01/13/versioning-in-the-libvirt-library/ That of course adds overhead on releases. nettle doesn't necessarily need to follow that process, I send that issue in order to spark discussion on whether that's desirable to have. regards, Nikos

3 6

What should nettle-3.5 be like?
by nisse＠lysator.liu.se 28 Mar '18

28 Mar '18

I used to think that next Nettle release should be focused on performance and add a few more features, no incompatible changes. But since the gcm changes breaks gnutls (fix in https://gitlab.com/gnutls/gnutls/commit/176aa191380e5d63a981f08f537deffdd78…), best way to reduce damage might be to bump the nettle soname. And if we do that, there are a couple of other things that could be done at the same time. Some changes to consider: 1. Remove the the symbols nettle_hashes and nettle_secp_384r1 and friends from the interface. They should be renamed with leading underscore. Important to be able to merge ed448 work without breaking the abi again. 2. Delete the old aes_* interface, in favor of aes128_, aes192_* and aes256_*. This could help aesni performance, where it might be nice with separately completely unrolled code for each key size. 3. Change struct nettle_aead to be more message-oriented. 4. Hide undocumented and internal symbols (the ones with leading underscore) more. Move declarations to internal uninstalled headers, possible tweak linker script to not expose at all in the shared library. 5. Revamp hmac and underlying hash functions with a separate state struct. Probably low priority, but it is a bit silly that, e.g., hmac_sha512_ctx includes three 128-byte large block buffers. Regards, /Niels -- Niels Möller. PGP-encrypted email is preferred. Keyid 368C6677. Internet email is subject to wholesale government surveillance.

5 22

Re: [PATCH 1/3] Add arm endianness-aware assembly infrastructure
by nisse＠lysator.liu.se 26 Mar '18

26 Mar '18

Michael Weiser <michael.weiser(a)gmx.de> writes: > Introduce m4 macros to conditionally handle differences of little- and > big-endian arm in assembler code. This and the next two patches now merged to master-updates for some final testing. At the moment, I'm only able to test that plain little-endian arm is still working. Thanks! /Niels -- Niels Möller. PGP-encrypted email is preferred. Keyid 368C6677. Internet email is subject to wholesale government surveillance.

2 1

mailing list archives
by Nikos Mavrogiannopoulos 20 Mar '18

20 Mar '18

Hi, On certain occasions I was trying to link to posts in that list however it seems that the official archives are not updated on a reasonable frequency. For example the current archives [0] were last updated on January 4. Is it possible to increase the archive generation frequency to weekly or sooner? regards, Nikos [0]. https://lists.lysator.liu.se/pipermail/nettle-bugs/2018/date.html

2 3

Efficient ED25519 on microcontrollers
by nisse＠lysator.liu.se 14 Mar '18

14 Mar '18

This is vaguely related to Nettle, but may be of interest to some people on the list. I've tried to summarize how to implement ed25519 efficiently on low-end hardware, with MSP430 as a concrete example. Most of the tricks are used by Nettle's implementation as well. Not sure if it would make any sense to try to port Nettle to such systems, but might be doable if there's interest in it. I would imagine that the need for good crypto on cheap devices is only going to increase in coming years. See https://www.lysator.liu.se/~nisse/misc/ed25519-msp430.html. Regards, /Niels -- Niels Möller. PGP-encrypted email is preferred. Keyid 368C6677. Internet email is subject to wholesale government surveillance.

1 0

x86 sha_ni
by nisse＠lysator.liu.se 13 Mar '18

13 Mar '18

Hi, I've been trying out the sha_ni instructions available on some newer x86_64 processors. Below replacement for sha1-compress.asm seems to run on roughly 2 cycles/byte when I benchmark it on an "AMD Ryzen 7 1700X" cpu in the gcc compile farm. Still sligthly slower than openssl, to squeeze out a few more cycles, it might help to change the sha1_compress interface to let it process more than one 64-byte block at a time. I hope to be able to wire it up via fat-x86_64.c reasonably soon. In the mean time, if anyone wants to try it out, just change the sha1-compress.asm symlink to point to this file. Regards, /Niels -----8<-------- C x86_64/sha_ni/sha1-compress.asm ifelse(< Copyright (C) 2018 Niels Möller This file is part of GNU Nettle. GNU Nettle is free software: you can redistribute it and/or modify it under the terms of either: * the GNU Lesser General Public License as published by the Free Software Foundation; either version 3 of the License, or (at your option) any later version. or * the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. or both in parallel, as here. GNU Nettle is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received copies of the GNU General Public License and the GNU Lesser General Public License along with this program. If not, see http://www.gnu.org/licenses/. >) C Register usage. C Arguments define(<STATE>,<%rdi>)dnl define(<INPUT>,<%rsi>)dnl define(<MSG0>,<%xmm0>) define(<MSG1>,<%xmm1>) define(<MSG2>,<%xmm2>) define(<MSG3>,<%xmm3>) define(<ABCD>,<%xmm4>) define(<E0>,<%xmm5>) define(<E1>,<%xmm6>) define(<ABCD_ORIG>, <%xmm7>) define(<E_ORIG>, <%xmm8>) define(<SWAP_MASK>,<%xmm9>) C QROUND(M0, M1, M2, M3, E0, E1, TYPE) define(<QROUND>, < sha1nexte $1, $5 movdqa ABCD, $6 sha1msg2 $1, $2 sha1rnds4 <$>$7, $5, ABCD sha1msg1 $1, $4 pxor $1, $3 >) .file "sha1-compress.asm" C _nettle_sha1_compress(uint32_t *state, uint8_t *input) .text ALIGN(16) .Lswap_mask: .byte 15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0 PROLOGUE(_nettle_sha1_compress) C save all registers that need to be saved W64_ENTRY(2, 10) movups (STATE), ABCD movd 16(STATE), E0 movups (INPUT), MSG0 movdqa .Lswap_mask(%rip), SWAP_MASK pshufd $0x1b, ABCD, ABCD pshufd $0x1b, E0, E0 movdqa ABCD, ABCD_ORIG movdqa E0, E_ORIG pshufb SWAP_MASK, MSG0 paddd MSG0, E0 movdqa ABCD, E1 sha1rnds4 $0, E0, ABCD C Rounds 0-3 movups 16(INPUT), MSG1 pshufb SWAP_MASK, MSG1 sha1nexte MSG1, E1 movdqa ABCD, E0 sha1rnds4 $0, E1, ABCD C Rounds 4-7 sha1msg1 MSG1, MSG0 movups 32(INPUT), MSG2 pshufb SWAP_MASK, MSG2 sha1nexte MSG2, E0 movdqa ABCD, E1 sha1rnds4 $0, E0, ABCD C Rounds 8-11 sha1msg1 MSG2, MSG1 pxor MSG2, MSG0 movups 48(INPUT), MSG3 pshufb SWAP_MASK, MSG3 QROUND(MSG3, MSG0, MSG1, MSG2, E1, E0, 0) C Rounds 12-15 QROUND(MSG0, MSG1, MSG2, MSG3, E0, E1, 0) C Rounds 16-19 QROUND(MSG1, MSG2, MSG3, MSG0, E1, E0, 1) C Rounds 20-23 QROUND(MSG2, MSG3, MSG0, MSG1, E0, E1, 1) C Rounds 24-27 QROUND(MSG3, MSG0, MSG1, MSG2, E1, E0, 1) C Rounds 28-31 QROUND(MSG0, MSG1, MSG2, MSG3, E0, E1, 1) C Rounds 32-35 QROUND(MSG1, MSG2, MSG3, MSG0, E1, E0, 1) C Rounds 36-39 QROUND(MSG2, MSG3, MSG0, MSG1, E0, E1, 2) C Rounds 40-43 QROUND(MSG3, MSG0, MSG1, MSG2, E1, E0, 2) C Rounds 44-47 QROUND(MSG0, MSG1, MSG2, MSG3, E0, E1, 2) C Rounds 48-51 QROUND(MSG1, MSG2, MSG3, MSG0, E1, E0, 2) C Rounds 52-55 QROUND(MSG2, MSG3, MSG0, MSG1, E0, E1, 2) C Rounds 56-59 QROUND(MSG3, MSG0, MSG1, MSG2, E1, E0, 3) C Rounds 60-63 QROUND(MSG0, MSG1, MSG2, MSG3, E0, E1, 3) C Rounds 64-67 sha1nexte MSG1, E1 movdqa ABCD, E0 sha1msg2 MSG1, MSG2 sha1rnds4 $3, E1, ABCD C Rounds 68-71 pxor MSG1, MSG3 sha1nexte MSG2, E0 movdqa ABCD, E1 sha1msg2 MSG2, MSG3 sha1rnds4 $3, E0, ABCD C Rounds 72-75 sha1nexte MSG3, E1 movdqa ABCD, E0 sha1rnds4 $3, E1, ABCD C Rounds 76-79 sha1nexte E_ORIG, E0 paddd ABCD_ORIG, ABCD pshufd $0x1b, ABCD, ABCD movups ABCD, (STATE) pshufd $0x1b, E0, E0 movd E0, 16(STATE) W64_EXIT(2, 10) ret EPILOGUE(_nettle_sha1_compress) -- Niels Möller. PGP-encrypted email is preferred. Keyid 368C6677. Internet email is subject to wholesale government surveillance.

3 9

Improved ECC tables
by nisse＠lysator.liu.se 10 Mar '18

10 Mar '18

Nettle uses precomputed tables for ecc scalar multiplication involving the fixed base point. I've found that parameters for those tables were somewhat poorly chosen. Updated parameters just pushed to branch ecc-params-tweak. The size of the tables is unchanged (except for the 192-bit curve, where they are reduced from 15 KB to 12 KB). In my benchmarking, 10%-15% speedup of ecdsa with the 256-, 384- and 521-bit curves, and 24% speedup of eddsa25519 (all for the signature operation, changes for the verify operation are more modest). I target table sizes of around 16 KB per curve. Would it be useful with a configure option to use smaller or larger table size? For the larger curves, one could perhaps get an additional 50% speedup with considerable larger curves. (But due to the requirements for side-channel silence, huge tables are impractical, since each access reads the entire table.) Regards, /Niels -- Niels Möller. PGP-encrypted email is preferred. Keyid 368C6677. Internet email is subject to wholesale government surveillance.

1 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

nettle-bugs March 2018