Received-SPF: pass (google.com: domain of linux-crypto+bounces-1421-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.48.161 as permitted sender) client-ip=147.75.48.161;
From: Dongsoo Lee <letrhee@nsr.re.kr>
To: 'David Laight' <David.Laight@ACULAB.COM>, 
	'Herbert Xu' <herbert@gondor.apana.org.au>, 
	"'David S. Miller'" <davem@davemloft.net>, 
	'Jens Axboe' <axboe@kernel.dk>, 'Eric Biggers' <ebiggers@kernel.org>, 
	"'Theodore Y. Ts'o'" <tytso@mit.edu>, 
	'Jaegeuk Kim' <jaegeuk@kernel.org>, 
	'Thomas Gleixner' <tglx@linutronix.de>, 
	'Ingo Molnar' <mingo@redhat.com>, 'Borislav Petkov' <bp@alien8.de>, 
	'Dave Hansen' <dave.hansen@linux.intel.com>, x86@kernel.org, 
	"'H. Peter Anvin'" <hpa@zytor.com>
Cc: linux-crypto@vger.kernel.org, linux-block@vger.kernel.org, 
	linux-fscrypt@vger.kernel.org, linux-kernel@vger.kernel.org
References: <20240112022859.2384-1-letrhee@nsr.re.kr> <20240112022859.2384-6-letrhee@nsr.re.kr> <cbd8de6ff70849a98faf2fd25b065a94@AcuMS.aculab.com>
In-Reply-To: <cbd8de6ff70849a98faf2fd25b065a94@AcuMS.aculab.com>
Subject: RE: [PATCH v6 RESEND 5/5] crypto: LEA block cipher x86_64 optimization
Date: Mon, 15 Jan 2024 14:17:43 +0900
Message-ID: <000e01da4772$2b2c5360$8184fa20$@nsr.re.kr>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain;
	charset="utf-8"
Content-Transfer-Encoding: quoted-printable
Content-Language: ko
Thread-Index: AQMGEr+nvC0x82l43QjPVTp1bc6i9wFH8CyOAqmBDhmuYzLfEA==

On  Fri, 12 Jan 2024 12:59:56 +0000, David Laight =
<David.Laight@ACULAB.COM> wrote:
> From: Dongsoo Lee
>> Sent: 12 January 2024 02:29
>>
>> For the x86_64 environment, we use AVX-512F/AVX2/SSE2 instructions.
>> Since LEA uses 128-bit blocks of four 32-bit integers, for =
optimization,
>> SSE2 encrypts 4 blocks, AVX2 encrypts 4/8 blocks, and AVX-512F =
encrypts
>> 4/8/16 blocks at a time.
>>
>> Our submission provides a optimized implementation of ECB, CBC
>> decryption, CTR, and XTS cipher operation modes on x86_64 CPUs
>> supporting.
>=20
> Given you say in 0/0:
>=20
> The LEA algorithm is a lightweight block cipher that processes data =
blocks of 128-bits and has three different key lengths, each with a =
different number of rounds:
>=20
> Just how big is it ?
> Doesn't look 'lightweight' to me.
>=20
> 	David
>=20
> -
> Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, =
MK1 1PT, UK
> Registration No: 1397386 (Wales)
>=20

Firstly, it's worth mentioning that LEA is an encryption algorithm =
designed to ensure 128-bit security.

The LEA cipher provides a balance between code size and required memory, =
allowing for trade-offs with performance. The implementation of LEA that =
we have submitted is oriented towards achieving optimal performance.

While it's difficult to compare to the most recent implementations of =
cryptographic algorithms because the test is out of date, you can see =
the results of a previous FELICS test [1] that implemented LEA for a =
smaller code size.

For example, a detailed example of skipping the key schedule to reduce =
memory usage and minimize code size can be found in [2].

Thank you for your interest.

    Dongsoo Lee

[1] https://www.cryptolux.org/index.php/FELICS#Results
[2] =
https://github.com/cryptolu/FELICS/blob/master/block_ciphers/source/ciphe=
rs/LEA_128_128_v03/source/encrypt.c