[pycrypto] PyCrypto AND Crypt_RSA integration

Thu Feb 12 21:47:15 CST 2009

Yes, I noticed that, I'm using 2048 in production, and I'll start using
urandom too. But that's not my problem, currently I'm focusing on other
things, so I'll just wait, meanwhile if someone find anything, it's all
good, otherwise when I need I'll ask for that other RSA implementation.

On Tue, Feb 10, 2009 at 10:32 PM, Dwayne C. Litzenberger <dlitz at dlitz.net>wrote:

> On Mon, Feb 09, 2009 at 10:36:40AM -0200, Mauricio Arozi wrote:
> >Am I helpless?
>
> I think the problem is that you're asking the mailing list for the *Python*
> Cryptography Toolkit about how to use an obscure *PHP* library.
>
> We can help with the Python side of things.  I wouldn't expect the people
> here to know and/or care much about PHP.
>
> > According to this site: http://pajhome.org.uk/crypt/rsa/rsa.html, and
> > yet others, the e(exponent?) is used for the public key, and d for the
> > private key.
>
> The notation I've seen most often is something like this:
>
>     n - modulus (public)
>     e - public exponent
>     d - private exponent
>     (n, e) - public key
>     (n, d) - private key
>     (p, q) - the (private) primes from which the keypair is derived.
>
> PyCrypto uses a similar notation:
>
>     from Crypto.PublicKey import RSA
>      import os
>
>     # DO NOT USE RandomPool (see below)
>     keypair = RSA.generate(2048, os.urandom)
>
>     print "PRIVATE KEYPAIR:"
>     print "n:", keypair.n   # modulus (public)
>     print "e:", keypair.e   # public exponent
>     print "d:", keypair.d   # private exponent
>     print "p:", keypair.p   # prime (private)
>     print "q:", keypair.q   # other prime (private)
>     print "u:", keypair.u   # I forget what this for (but it's private)
>
>     pub = keypair.publickey()
>     print ""
>     print "PUBLIC KEY:"
>     print "n (pub):", pub.n     # modulus (public)
>     print "e (pub):", pub.e     # public exponent
>     print "d (pub):", pub.d     # raises an exception
>     print "p (pub):", pub.p     # raises an exception
>     print "q (pub):", pub.q     # raises an exception
>     print "u (pub):", pub.u     # raises an exception
>
> This outputs the following:
>
>     PRIVATE KEYPAIR:
>     n: 277...[truncated]
>     e: 65537
>     d: 232...[truncated]
>     p: 159...[truncated]
>     q: 174...[truncated]
>     u: 125...[truncated]
>
>     PUBLIC KEY:
>     n (pub): 277...[truncated]
>     e (pub): 65537
>     d (pub):
>     Traceback (most recent call last):
>       File "x.py", line 21, in ?
>         print "d (pub):", pub.d
>       File "/usr/lib/python2.4/site-packages/Crypto/PublicKey/RSA.py", line
> 154, in __getattr__
>         return getattr(self.key, attr)
>     AttributeError: rsaKey instance has no attribute 'd'
>
> > My problem is that while using PyCrypto to generate both public and
> > private keys, the e(exponent?) is always the same.
>
> Mads Kiilerich already talked a bit about this, but I won't go into detail.
> What you're describing here is normal, and it really helps improve the
> performance of encryption/verification.
>
> If you're concerned about the security of using RSA, I suggest reading Dan
> Boneh's 1999 article, "Twenty years of attacks on the RSA cryptosystem":
>
>     http://crypto.stanford.edu/~dabo/abstracts/RSAattack-survey.html<http://crypto.stanford.edu/%7Edabo/abstracts/RSAattack-survey.html>
>
> >So in simple words, I only need to be able to encrypt/decrypt sign and
> >verify signs on php and python, simultaneously, if possible, using RSA
> >algo.
>
> PyCrypto's PublicKey package is very low-level, so people shouldn't use it
> directly unless they REALLY know what they are doing.  Mere mortals should
> use a separate library in addition to PyCrypto for that.  You should never
> do anything like this:
>
> >privkeyA = RSA.generate(512, rpool.get_bytes)
> >pubkeyA = privkeyA.publickey()
> >
> >msg = 'This is the secret phrase testing.'
> >msgc = pubkeyA.encrypt(msg, '')
>
> That is called "textbook RSA", and it's insecure.  (Also, it uses a 512-bit
> key, which is way too short, but I assume that's just for demonstration.)
> I strongly recommend looking at PKCS#1v2 (also known as RSAES-OAEP).
> PyCrypto doesn't include an implementation yet, but Sergey Chernov
> mentioned that he is working on one.
>
> Also, I noticed in your code that you used RandomPool.  Don't.  RandomPool
> is a security disaster, and it will be removed from future versions.  See
> the following messages:
>
>     http://lists.dlitz.net/pipermail/pycrypto/2008q3/000000.html
>     http://lists.dlitz.net/pipermail/pycrypto/2008q3/000020.html
>
> I hope you find the above information helpful.
>
> Cheers,
>  - Dwayne
>
> --
> Dwayne C. Litzenberger <dlitz at dlitz.net>
>  Key-signing key   - 19E1 1FE8 B3CF F273 ED17  4A24 928C EC13 39C2 5CF7
>  Annual key (2008) - 4B2A FD82 FC7D 9E38 38D9  179F 1C11 B877 E780 4B45
> _______________________________________________
> pycrypto mailing list
> pycrypto at lists.dlitz.net
> http://lists.dlitz.net/cgi-bin/mailman/listinfo/pycrypto
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.dlitz.net/pipermail/pycrypto/attachments/20090213/d009e7ea/attachment-0001.htm 