[security-dev 00891]: Re: code review request 6570344 Invalid RSA OID in sun.security.x509.AlgorithmId

[Xuelei Fan]

Thanks for the review.

I have though all certificate should adopt the PKCS#1, i.e, using the 
algorithm OID, 1.2.840.113549.1.1.1. But I did find a certificate just a 
few minutes ago that still use the OID, 2.5.8.1.1. Windows and Firefox 
recognized the certificate.

With the fix, the certificate could be validated. Otherwise, thrown a 
exception "java.security.InvalidKeyException: Not an RSA key: 2.5.8.1.1".

But by now, we did not get complains on the above exception, it is 
really really a very rare use case.

Thanks,
Xuelei

Sean Mullan wrote:
> The fix looks fine to me.
>
> --Sean
>
> Xuelei Fan wrote:
>> Hi,
>>
>> The RSA OID from sun.security.x509.AlgorithmId is 1.2.5.8.1.1. 
>> However no such OID seems to exist. The correct one should be 2.5.8.1.1.
>>
>> ITU-T X.509 defined RSA encryption algorithm as:
>> id-ea-rsa = {joint-iso-itu-t(2) ds(5) algorithm(8) 
>> encryptionAlgorithm(1) rsa(1)}
>> rsa ALGORITHM ::= {
>>    KeySize
>>    IDENTIFIED BY id-ea-rsa
>> }
>>
>> However, the industry does not use the above specification, a serial 
>> of definitions of PKCS#1 are adopted instead (the PKIX WG of IETF 
>> adopts the PKCS#1 definitions). I think that is also why we did not 
>> get issue report on parsing a certificate with such a OID. BTW there 
>> is a defect report to deprecate the above definition. [1]
>>
>> Anyway, I think we need to correct "1.2.5.8.1.1" to "2.5.8.1.1" even 
>> no practical certificate issues reported by now.
>>
>> Webrev: http://cr.openjdk.java.net/~xuelei/6570344/webrev.00/
>> Bug description: http://cr.openjdk.java.net/~xuelei/6570344/webrev.00/
>>
>> [1]: 
>> http://www.oid-info.com/cgi-bin/display?oid=2.5.8.1.1&submit=Display&action=display 
>>
>>
>>
>> Thanks,
>> Xuelei
>