Re: OpenSSL RSA 1024 bits implementation broken?

[Dave Horsfall <dave () horsfall org>]

On Mon, 6 Oct 2014, Pierre Schweitzer wrote:

> There appear to have some noise on the Internet regarding a possible 
> flaw in the 1024 bits RSA implementation in OpenSSL which would allow 
> bruteforcing the private key in ~20 minutes.
>
> Does anyone has any information about this? The associated pastebin to 
> the said information is: http://pastebin.com/D8itq6Ff Is this serious?

On the moderated crypto list where I hang out, it's receiving much 
attention.  The consensus is that it's likely a buggy compiler or 
optimiser that rounded integer division upwards instead of truncating it 
as required by the C standard, and that the "discoverer", by refusing to 
provide further details, is full of it.

You may be able to search the archives at cryptography () metzdowd com; as I 
said it's a moderated list, but full of techie people who really know 
their onions.

-- Dave