[
https://issues.apache.org/jira/browse/WSS-300?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13072658#comment-13072658
]
Linton Miller commented on WSS-300:
-----------------------------------
Unfortunately, the new SKITest test case doesn't test this fix. That's because
the test certificate has the SKI extension, thus the getSKIBytesFromCert
doesn't need to calculate it from the public key (using the new code), but
simply reads the extension value. You need to generate test certificates that
don't have the X509v3 Subject Key Identifier extension (like my attached
example) to test this new code.
> SubjectKeyIidentifier (SKI) incorrectly calculated for 2048-bit RSA key
> -----------------------------------------------------------------------
>
> Key: WSS-300
> URL: https://issues.apache.org/jira/browse/WSS-300
> Project: WSS4J
> Issue Type: Bug
> Components: WSS4J Core
> Affects Versions: 1.5.11, 1.6.1
> Environment: Tomcat 5, Solaris 10, Java 1.4 and Tomcat 6, Win XP,
> Java 6
> Reporter: Linton Miller
> Assignee: Colm O hEigeartaigh
> Fix For: 1.5.12, 1.6.2
>
> Attachments: CryptoBase.java.patch, DERDecoder.java,
> X509SubjectPublicKeyInfo.java, X509SubjectPublicKeyInfo.java, broken.xml,
> client.wsdd, correct.xml, rsa2048.jks, test.cypto.properties
>
>
> The crypto function to get the SubjectKeyIdentifier from an X509Certificate
> has incorrect hard-coded assumptions about the size of the encoded
> information, meaning the calculation of the SKI from a 2048-bit RSA key is
> incorrect.
> The method
> org.apache.ws.security.components.crypto.CryptoBase.getSKIBytesFromCert does
> not parse the DER encoding of information, but just tries to pick out the
> piece of the byte array that corresponds to the content of interest. However,
> that approach fails because the DER encoding is variable length, depending on
> the size of the data being encoded. e.g. a 1024-bit key in a DER BIT STRING
> takes 4 bytes header + 140 bytes data to encode, whereas a 2048-bit key
> takes 5 bytes header + 270 bytes data; the header is one byte longer for the
> larger key, so the data starts at a different point in the array.
> To fix this, the DER data structures should be properly processed, reading
> the DER header bytes to determine the length of each data element (that also
> allows the generalization of handling any X.509 encoded public key, rather
> than just RSA keys as currently coded).
> Attached is a suggested patch (against WSS4J 1.6.1) that implements this
> idea: it processes the SubjectPublicKeyInfo and SubjectKeyIdentifier
> DER-encoded byte arrays according to their ASN.1 definitions from RFC
> 3280/5280 to pick out the desired data bytes.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@ws.apache.org
For additional commands, e-mail: dev-h...@ws.apache.org
