Date: Tue, 12 Apr 2005 18:35:49 +0200
From: Eric Rannaud <Eric.Rannaud@ens.fr>
To: Pedro Larroy <piotr@larroy.com>
Cc: linux-kernel@vger.kernel.org
Subject: Re: Call to atention about using hash functions as content indexers (SCM saga)
Message-ID: <20050412163549.GA7379@clipper.ens.fr>
References: <20050411224021.GA25106@larroy.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20050411224021.GA25106@larroy.com>
User-Agent: Mutt/1.4.2i
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1040
Lines: 24


Simply put, the best known attack of SHA-1 takes 2^69 hash operations.
( http://www.schneier.com/blog/archives/2005/02/sha1_broken.html )
The attack is still only an unpublished paper and has not yet been
implemented. An attack is: you try as hard as you can to find a collision
between two arbitrary messages (i.e. two arbitrary --and nonsensical--
source files).
In the context of git, a better estimation would be the number of hash
operations needed to find a message that has the same hash than a given
fixed message (e.g. mm/memory.c). This is more like 2^100 hash
operations. And if a collision is found, this is very likely using a
message that *doesn't* look like a C source file...

Moreover, no example of collision is known, AFAIK.

In other words: this won't happen.

Best,
   /er.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/