netdev - Re: [PATCH cryptodev] crypto: lib/chacha20poly1305

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 linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

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

Message-ID: <69c23c36-60fe-2676-d07a-67ce1f752dd1@unstable.cc>
Date:   Tue, 17 Nov 2020 10:41:10 +0100
From:   Antonio Quartulli <a@...table.cc>
To:     "Jason A. Donenfeld" <Jason@...c4.com>
Cc:     Linux Crypto Mailing List <linux-crypto@...r.kernel.org>,
        WireGuard mailing list <wireguard@...ts.zx2c4.com>,
        Netdev <netdev@...r.kernel.org>, keyrings@...r.kernel.org,
        linux-security-module <linux-security-module@...r.kernel.org>,
        Antonio Quartulli <antonio@...nvpn.net>,
        Herbert Xu <herbert@...dor.apana.org.au>,
        David Howells <dhowells@...hat.com>,
        Jarkko Sakkinen <jarkko@...nel.org>,
        Jakub Kicinski <kuba@...nel.org>
Subject: Re: [PATCH cryptodev] crypto: lib/chacha20poly1305 - allow users to
 specify 96bit nonce

Hello,

On 17/11/2020 09:30, Jason A. Donenfeld wrote:
> Nack.
> 
> This API is meant to take simple integers, so that programmers can use
> atomic64_t with it and have safe nonces. I'm also interested in
> preserving the API's ability to safely encrypt more than 4 gigs of
> data at once. Passing a buffer also encourages people to use
> randomized nonces, which isn't really safe. Finally, there are no
> in-tree users of 96bit nonces for this interface. If you're after a
> cornucopia of compatibility primitives, the ipsec stuff might be more
> to your fitting. Or, add a new simple function/api. But adding
> complexity to users of the existing one and confusing future users of
> it is a non-starter. It's supposed to be deliberately non-awful to
> use.
> 

Thanks for explaining the ratio behind this API.

At first I thought this API wanted to take over the existing one, hence
my attempt of making it more generic and re-use it.
But I understand now this was not the goal.

I will stick to the classic crypto API then.

Best Regards,


p.s. I am curious about any use case you may have in mind for encrypting
more than 4GB in one go, as there are no users doing that right now.

-- 
Antonio Quartulli