phc-discussions - Re: [PHC] Argon2

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Wed, 08 Apr 2015 05:40:24 +0300
From: Alexander Cherepanov <ch3root@...nwall.com>
To: discussions@...sword-hashing.net
Subject: Re: [PHC] Argon2

On 2015-03-31 18:35, Solar Designer wrote:
> On Tue, Mar 31, 2015 at 03:04:43PM +0100, Samuel Neves wrote:
>> On 31-03-2015 14:30, Solar Designer wrote:
>>>>>          blake2b_update(&BlakeHash, (const uint8_t*)&lanes, sizeof(lanes));
>>> [...]
>>>
>>> Surely blake2b_update() then uses its own choice of integer or SIMD
>>> types for this data.
>>
>> The BLAKE2 code does not treat input bytes as anything other than bytes. Whenever there is a need to convert bytes to
>> integers, this is done in a standards-compliant way (using either memcpy to a local variable, or plain load-shift-or).
>
> Thanks!  Yes, blake2b_update() appears to always hit a memcpy() on "in".
>
> Perhaps that's the way to write such functions these days, although there
> must be a performance hit from this when running on an architecture such
> as x86 that would permit you to access integers in "in" directly.

According to http://blog.regehr.org/archives/959 , memcpy could be 
optimized away.

-- 
Alexander Cherepanov